Peter Reviews Products & Procedures

Bosch S 800 Linea: Simple Black Out Fix

2012-01-22T12:04:00.000-06:00

Roughly 18 months ago, we bought this shiny marvel of German über-engineering also known as 36" Counter-Depth Side-by-Side Refrigerator 800 Series - Stainless Steel (B22CS80SNS). The website slogan proclaims “Invented for Life.” The awesome appliance is absolutely invented for the space age with 21^st-century amenities and countless energy-saving features like low-power consuming light fixtures in the backlit refrigerating compartment. Soon after installation, however, the fantastic illumination began to flicker on opening the door and eventually dropped out entirely.

Our first idea from decades of life experience with refrigerators was that some light bulb needed changing. After thorough inspection, however, no conventional light bulbs were to be found. Under the ceiling cover, we found two tiny LED lights that we first did not recognize as light sources at all and do not look like they should be removed. There is no easy way to open the cover of the back illumination either. The back is lit by a Laser light source. The manual recommends to call a service technician. Welcome to German precision goods in the 21^st century! No way for less-talented rocket scientists to even change a light bulb!

But before we extend a time-consuming and costly invitation to Herr Dipl. Ing. Bosch's faithful associates, here is a remedy easy enough for us to try. When we open the door to the refrigerator compartment, we see two push button switches at the top of the refrigerator's frame. The left one is a small rod that is pushed in by the door when we close it. The right switch shows a reboot symbol adopted from the PC world. That switch represents the on/off switch for the whole refrigerator. When we pushed it, the lighting in the compartment - ceiling and back - sprang to life briefly, before the refrigerator shut down and the display above the water and ice dispenser fell dark. We pushed the switch again to restart the refrigerator.

The short resurrection of the compartment lighting during shutdown suggested that the light sources were actually still intact, but were not switched properly.

The video above shows the push-rod switch in action.

This observation leads us to the function of the push-rod switch on the left. The rod is spring-loaded. That is, the switch turns the light off when the door closes, holding the rod down, and is supposed to switch the light back on when the door is opened and the rod is depressed. We hypothesized that perhaps the switch mechanism was stuck and the spring did not release it when the rod was depressed. To test this idea, we squirted a tiny amount of light spray oil like WD-40^® into the small gap between the rod and the switch cover, and sure enough our compartment light is back. Enjoy this otherwise fantastic fridge and good luck with aiming the ice cubes!

Seismic Activity & Reactor Safety: Lessons from Japan

2011-10-27T19:16:00.000-05:00

I worked with radioactive isotopes known as radionuclides in laboratory research for almost three decades in my professional career. Since three reactor fuel cores have melted down at the Fukushima Dai-ichi (number one) Nuclear Power Station (NPS) owned by Tokyo Electric Power Company (TEPCO) in the aftermath of the Tohoku-Chihou-Taiheiyou-Oki Earthquake and Tsunami on March 11, 2011 [1], I am concerned about the safety of nuclear power reactors.

At magnitude M_w 9.0 (I_JMA 7.0 on the Japan Meteorological Agency intensity scale), this earthquake was the strongest recorded in Japan's history, causing massive damage. Moreover, tsunami waves up to 40 meters high inflicted horrible devastation on the coast. The Government of Japan reports that 15,687 people lost their lives and 4,757 are still missing as of Aug. 9, 2011 [2].

The quake and tsunami caused a total station blackout at Fukushima Dai-ichi NPS. Reactor fuel core cooling could not be maintained, the fuel began to melt, and radioactive material could not be contained. After powerful hydrogen explosions, still unknown amounts of radioactive material were released into the air and the sea. Roughly 80,000 people living in immediate proximity of the power station were evacuated and have not been able to return home, except for brief visits [2]. Soil and sea near the NPS are highly contaminated with radioactive material, and long-lived radionuclides have been found foods like rice [3] and beef [4] at levels too high for consumption. Hotspots of radiation are being discovered in Tokyo and further away [5]. The impact on public health is not yet fully comprehended. The costs of this crisis and its long-term consequences are unfathomable, but will be in the hundreds of billion yen.

Aerial photograph of Fukushima 1 (Dai-ichi) NPS taken March 24, 2011. Note the obliteration of the reactor buildings at Units 1 (top), 3 and 4. Fuel cores melted in Units 1, 2 and 3 because of loss of coolant. Unit 4 was shutdown for service at the time of the earthquake and tsunami (source cryptome.org).

Japan lies in a seismically active zone near the Pacific Rim of Fire. The islands are situated in close proximity to the Japan Trench in the Pacific Ocean off Japan's Northeast Coast where two tectonic plates violently collide. No doubt Japanese engineers have collected in-depth theoretical and empirical knowledge on the effects of seismic waves on industrial systems, structures and components, particularly in nuclear engineering. All nuclear power stations in Japan are equipped with seismometers that record ground motion in three dimensions and automatically scram nuclear reactors at predetermined set-points.

Aerial photograph of Kashiwasaki-Kariwa NPS. Older Units 1 (bottom, right) to 4 (top) are in the foreground. The most recently built Units 5 to 7 are in the background (top, left) (source TEPCO).

Kashiwasaki-Kariwa NPS
As recently as Jul. 16, 2007, a M_w 6.6-magnitude (I_JMA& 6.8) earthquake known as the Niigata-ken Chuetsu-Oki Earthquake, or NCO Earthquake for short, struck Kashiwazaki-Kariwa Nuclear Power Station (KK) on Japans Northwest coast. The NPS, with seven reactors and 8.2 Gigawatt electrical power output the world's largest, is owned by the same utility as Fukushima Dai-ichi NPS known as TEPCO. The quake automatically tripped the four reactors that were operating. Kayen and others (2007) [6] examined the geotectal causes for the quake in great detail and compiled a comprehensive account on land failure and the damage to infrastructure in the region. I have explained fundamental wave concepts underlying earthquakes in the context of the Great Haiti Earthquake of Jan. 12, 2010, in my post with the title "Neuroanatomy of an Earthquake" published Jan. 18 of that year. In this essay, I shall attempt to explain in more detail measurements of ground motion, projections of observations, and their relationship with quake damage at nuclear power stations.

Ground Motion Acceleration Units, Dimensions, and Conversions

1 Gal (Galileo) = 1 cm/s²

1 g (fraction of gravitational acceleration) = 980 Gal

At Kashiwasaki-Kariwa NPS, operators were able to recover recordings of ground motion from only one third of roughly 100 recording locations [7]. Fortunately, recordings could be retrieved from the bottom floors of the seven reactor buildings, known as base mats. As typical examples, the accelerogram for the East-West ground motion recorded on the base mat of Kashiwasaki-Kariwa NPS Unit 2 is shown below as well as the observed acceleration response spectrum (thick line) and the design-basis (S2, thin line) acceleration response spectrum for the same location. The latter is computed, using synthesized time histories of passed earthquakes.

TEPCO [8]

Peak ground acceleration (PGA) on the reactor building base mats exceeded the design basis for all seven units. The greatest PGA was recorded at Kashiwasaki-Kariwa NPS Unit 1. The worst exceedance was 3.6-times design basis at Kashiwasaki-Kariwa NPS Unit 2 [8]. Note 5% damping is assumed in the calculations, because the structures absorb quake energy [9].

TEPCO [8]

The trip points for automatic reactor shutdown known as SCRAM were set at 100 Gal (0.10 g) in the horizontal directions and at 120 Gal (0.12 g) in the vertical direction [10].

The quake caused a massive switch yard transformer fire. Radioactive gaseous effluent was released into the atmosphere because of a malfunction of the HVAC system. The amount was deemed small enough not to pose a threat to the public. Visual inspection during extensive post-quake walkdowns, pressure and function tests were carried out on about 20,600 components, instruments, panels and valves and 155,000 meters of pipe [10]. Nuclear power station systems, structures and components (SSC) are classified in three groups of importance for seismic safety. SSCs in the most important group did not appear to have incurred any damage at Kashiwasaki-Kariwa NPS. This group is known as class A in Japan and seismic design category I, or SDC-I for short, in the U.S (Antaki and Johnson, 2011) [11]. Damage, deformation, and impeded function were uncovered only on class B (SDC-II) and C (SDC-III) systems, structures and components deemed unimportant to reactor safety.

Using advanced ultrasonic technologies, inspectors detected some hidden damage. For example, cracks were found in the rotor shaft and blades of power-generating low pressure main steam turbines. It was concluded that this damage was a result of fatigue rather than the quake. All SSCs at which damage was discovered were repaired. Because of the broad exceedance of the design basis earthquake at the seven units, additional safeguards were installed, and new spectra were computed to ensure that the implemented measures strengthened the design basis.

The apparent disassociation between earthquake magnitude, observed peak ground acceleration and damage led to the search for new indicators better predicting quake damage. The currently used measurements of magnitude and ground motion do not adequately appreciate the temblor's duration. To remedy this shortcoming, cumulative absolute velocity expressed in cm/s or g-sec, CAV for short, has been introduced. CAV constitutes the sum of the ground accelerations added over the duration of the quake. Standardized cumulative absolute velocities are filtered for non-damaging ground motion frequencies.

Standardized cumulative absolute velocities [g-sec] are plotted against the magnitude of quakes that shook nuclear power stations before 2011. Magnitude is expressed as seismic intensity on the I_JMA scale used by the Japan Meteorological Agency (JMA). Historically, trained JMA agents qualitatively scored quake magnitude. Today, instruments register quake intensity automatically. Damage was detected at main turbines, that is SSCs unimportant to reactor safety. The data suggest a damage threshold for such SSCs at I_JMA 5.5 [12]. The lowest PGA the JMA allows at this intensity is 0.25 g. No quake damage is anticipated for PGAs below 0.25 g.

Foreshadowing the reactor catastrophe at Fukushima, however, in tests during the restart of Kashiwasaki-Kariwa Unit 7 in May 2009 the Reactor Core Isolation Cooling system, or RCIC for short, did not perform as expected because of valve failure. RCIC steam entered a pressure control chamber at the bottom of the reactor pressure vessel, and the cooling water in the chamber rose. Operators had to intervene. At Fukushima Dai-ichi NPS, the RCIC systems of the two operating reactors outfitted with this system would fail last March, profoundly impeding the operator's efforts of preventing the reactor fuel cores from melting down. I have described the weaknesses of the RCIC system in detail in my post with the title "Fukushima: Failure by Design" published Jul. 3, 2011.

TEPCO [13]

Fukushima Dai-ichi NPS
TEPCO insists that the total station blackout, caused by the loss of emergency diesel power through tsunami inundation, precipitated the reactor fuel core meltdowns at Fukushima Dai-ichi NPS [13]. The operator has given seismic impact as cause little credence. TEPCO's interpretation is supported by the ground motion recorded on the base mats of the reactor buildings. Though PGAs broadly exceeded design basis in the horizontal East-West direction at Fukushima Dai-ichi NPS, the greatest exceedance was roughly a third of that found at Kashiwasaki-Kariwa NPS after the NCO Earthquake, that is the greatest exceedance was only 1.26-fold at Unit 2.

Aerial photograph of Fukushima 1 (Dai-ichi) NPS taken Sep. 18, 2010. Units 4 (left) to 1 (right) were built in a row. Unit 1 closest to the water front is the oldest commercial nuclear power reactor of Japan completed in 1971. Units 5 and 6, set apart on the upper right, are the youngest reactors (source cryptome.org).

Of the six reactors at Fukushima Dai-ichi NPS, units 5 and 6 were least affected by the tsunami and exhibited the smallest PGA exceedance [13]. They were disconnected from the power grid when the quake struck, undergoing tests fully loaded with fuel. Combined with the findings at Kashiwasaki-Kariwa NPS after the NCO Earthquake, the seismic impact uncovered at these units may provide valuable lessons for predictions of future quake damage at nuclear power stations.

In fact, the lessons learned from both earthquakes in Japan stipulate that ground motion in exceedance of the design basis may impact nuclear power station systems, structures and components (SSC) in subtle ways. Although quake damage may mainly be detected on less quake-resistant SSC categorized B (SBC-II) and C (SBC-III), they warrant in-depth inspections for potential hidden damage of all SCC, regardless of class.

North Anna NPS
On Aug. 23, 2011, the Northeastern Seaboard of the U.S. was struck by a magnitude M_w 5.9 earthquake centered near Mineral, Virginia, in the Central Virginia Seismic Zone and in close proximity to the North Anna Nuclear Power Station with two pressurized water reactors.

USGS [14]

Both reactors were shut down safely. The quake was felt as far as Boston, Massachusetts [15]. By and large minor damage was found at monuments, as well as at governmental, residential and commercial buildings, extending over a surprisingly large area. This Washington Post picture gallery with the title "Magnitude-5.8 earthquake shakes D.C." provides a sweeping impression of the quake's impact.

The damage combined with subjective experience, matches the criteria of a six-lower (6弱) magnitude quake on the Japan Meteorological Agency (JMA) shindo scale, pegging the quake's PGA between 0.250 g and 0.315 g. In accord, the United States Geological Service (USGS) estimates 0.26 g [14].

According to the findings in Japan discussed above, this peak ground acceleration merely touches the threshold at which damage to systems, structures and components must be expected at nuclear power stations. Based on observations after the 1995 South Hyogo Prefecture Earthquake, also known as the Kobe Earthquake, Ochiai and others (2010) suggest that the potential damage at nuclear power plants with ground shaking of the magnitude of the Virginia quake may comprise: falling air ducts, tumble (with weak anchorage), failure of foundation anchorage, circular storage tank wall buckling (elephant foot), contact/hitting of pipes (insulation and grating damage), and buckling of crane basements. On occasion, pulling out/fracture of anchor bolts, overflow (sloshing), pipe support structure damage (pulling out of anchor bolts), and failure of transmission line support may be observed [12].

North Anna NPS is the first nuclear power station in the Central Northeastern U.S. at which reactors scrammed because of a seismic event. The reactors at North Anna NPS lack automatic seismic scram systems. The reasons for the scrams are not yet fully understood. Offsite power was lost because of ground motion. Emergency diesel generators started up. The ensuing power fluctuations may have tripped the reactors. Seismic recordings were obtained from the Unit 1 reactor containment base mat as well as the turbine buildings and sent out for evaluation in the days after the incident by the station's operator, (Dominion) Virginia Electrical and Power Company, or VEPCO for short. In addition, VEPCO immediately carried out walkdowns, visually inspecting the station's systems, structures and components (SSC) for damage. Operability and performance was examined extensively. Underground piping was unearthed and checked, and pressure-tested. Welds were examined with ultrasonic devices.
In public meetings at Nuclear Regulatory Commission (NRC) headquarters near Washington, D.C., on Sep. 8 and Oct. 21, 2011, VEPCO reported preliminary findings. VEPCO is determined to restart the reactors as soon as the inspections are completed, and the NRC approved the results.

Webcast of the September 8, 2011, NRC meeting with VEPCO on the Aug. 23 Virginia Earthquake near North Anna NPS (Docket Nos. 50-338 and 50-339; slides).

Webcast of the October 21, 2011, NRC meeting with VEPCO on the Aug. 23 Virginia Earthquake near North Anna NPS (slides).

At the Oct. 21 meeting, VEPCO presented accelerograms retrieved from the reactor building base mat of North Anna NPS Unit 1. Note the short duration of the quake. The figure below as well as other North Anna NPS data shown in this post were released for the two NRC meetings.

Accelerograms showing ground motion on the North Anna NPS Unit 1 reactor building base mat in the three spatial directions (left ordinate: acceleration [Gal]; right ordinate: [g], abscissa: time [s]). Note the ground motion at the NPS lasted only 3.1 seconds (VEPCO, Oct. 21, 2011).

In addition, VEPCO provided the peak ground accelerations (PGA) the company used to calculate the design basis spectra for North Anna NPS.

Stipulated peak ground acceleration limits for rock (reactor building foundation) and soil (foundation of turbine buildings and ancillary structures) by direction of the ground motion for Operating Basis Earthquake and Design Basis Earthquake (VEPCO, Sep. 9, 2011).

The stipulated PGAs vary for ground motion direction and ground type, that is rock and soil on which the reactor buildings and ancillary structures are built, respectively. According to Antaki and Johnson (2011)[11], NPS systems, structures and components are designed to withstand ground motion based on the operation basis earthquake, or OBE for short, which is anchored at half the PGA used to determine the Safe Shutdown Earthquake (SSE), that is at the ground motion at which the reactors are to be shut down to protect their safety. By contrast, the design basis earthquake, or DBE for short, is calculated to envelope the expected ground motion envisaged at 80 percent of the NPS sites in the Central and Eastern U.S., assuming an OBE equal to one-third or less of the SSE. Depending on NPS location, the PGA used to determine the SSE is set between 0.1 g and 0.3 g, which is equal to or higher than the trip points for automatic seismic scram systems used at Japanese nuclear reactors. In sum, the reactors are supposed to be shut down at peak ground accelerations much below DBE. However, note that the greatest anticipated DBE at North Anna NPS is 0.18 g. which is below the USGS estimate for the Virginia quake. Therefore, it does not come as a surprise that recorded peak ground motions were greater than assumed in the DBE.

Recorded (ondulating curves) and computed (smooth curves) seismic response spectra at 5 % percent damping for the vertical (top) and horizontal (bottom) directions at North Anna NPS Unit 1 reactor building base mat (ordinate: acceleration [g]; abscissa: frequency [Hz]). The bottom and middle smooth curves in each graph describe OBE and DBE, respectively. Note that the recorded ground motion exceeded DBE at frequencies greater than 1.0 Hz in the vertical as well as in the North-South direction, and greater than 10.0 Hz in the East-West direction (VEPCO, Sep. 9, 2011).

The two graphs above show the seismic response spectra constructed from the data VEPCO recovered from the base mat seismometers of the North Anna NPS Unit 1 reactor building. Clearly, PGA exceeded DBE at frequencies greater than one Hertz. Damage was to be expected at class B and C systems, structures, and components, that is SSCs unimportant for reactor safety.

However, VEPCO reasoned that the anticipated damage might not have been detrimental to the safety of North Anna NPS, because quake duration and DBE exceedance were short in comparison with other known damaging quakes. The operator reasoned that cumulative absolute velocity (CAV), which takes the duration of the quake into account, represents a more accurate damage predictor than PGA, as discussed in context with the NCO Earthquake and the Tohoku Earthquake. VEPCO therefore presented standardized CAV values at the NRC meetings.

Standardized cumulative absolute velocities [g-sec] calculated by three expert companies from the observed North Anna NPS Unit 1 reactor building base mat accelerations for ground motion in the three directions. Note that North-South direction motion exceeded the NRC prescribed limit of 0.16 g-sec below which a nuclear power reactor can be operated safely (VEPCO, Sep. 9, 2011).

Standardized cumulative absolute velocities [g-sec] calculated from recordings and modeled for the three directions of motion on the North Anna NPS Unit 1 reactor building base mat. The observed velocities did not exceed the modeled design base earthquake (DBE) and the limits modeled in the Individual Plant Examination for External Events (IPEEE) (VEPCO, Oct. 20, 2011).

The standardized cumulative absolute velocities determined for recorded North-South ground motion on the base mat of North Anna NPS Unit 1 reactor building exceeded OBE, necessitating the shutdown of both units. Yet, the values remained below the cumulative absolute velocities calculated for DBE as well as the higher limits calculated in the Individual Plant Examination for External Events (IPEEE) stipulated by the NRC. The IPEEE, but not the DBE values, are above 0.8 g-sec, that is the CAV threshold for quake damage at nuclear power stations in Japan, according to the CAV/quake magnitude plot of Ochiaia and others (2011)[12]. The authors note only one incident with damage that occurred at a CAV below 1.5 g-sec. Main steam turbine damage was detected at Hamaoka NPS Unit 5 after the Aug. 11, 2009, Shizuoka Earthquake, which scrammed the station's units 5 and 6.

By contrast, the greatest recording-based CAV at Kashiwasaki-Kariwa NPS reached 2.8 g-sec, 16-times the highest CAV at North Anna NPS. For ground motion predictions, Campbell and Bozorgnia (2011) plotted the standardized cumulative absolute velocity against I_JMA, that is the quake intensity the Japan Meteorological Agency (JMA) uses (Fig. 4.1) [16]. The plot shows a close direct proportionality between the logarithm of cumulative absolute velocity and the JMA quake intensity scale. VEPCO's cumulative absolute velocities for recorded ground motion fall short of the range between 0.3 and 2.0 g-sec (median = 1.2 g-sec) that the plot predicts for the Virginia quake, though the qualitative descriptions of the damage and the quake experience in the region closely match the ones provided by JMA for a quake of this magnitude. Therefore, VEPCO's cumulative absolute velocities may underrate the possible extent of the damage.

North Anna NPS lost offside power because of a disruption in the switchyard, necessitating the startup of emergency diesel generators. This event arguably represents the most consequential impact of the Virginia quake on the NPS known to date. So far, the station inspection uncovered minor damage to SDC-II and SDC-III systems, structures, and components unimportant to reactor safety. Intriguingly, the inspectors noted quake impact similar to that found at Fukushima Dai-ichi Units 5 and 6, though the recorded peak ground acceleration at North Anna NPS was only roughly half of that observed at these units at Fukushima Dai-ichi NPS, and, according to the Japan Meteorological Agency, the Tohoku Earthquake reached six-upper (6強) on the shindo scale with seismic intensities I_JMA ranging from 6.0 to 6.4 in the region of Fukushima Dai-ichi NPS, whereas the Virginia Earthquake is classified six-lower (6弱). As examples for similar quake impact, two types of visible damage are explored below.

Component Relocation: At Fukushima Dai-ichi Unit 5 turbine building a support for the moisture separator drain pipe shifted,

Moisture separator drain pipe support moved in Fukushima Dai-ichi Unit 5 turbine building (TEPCO).

while at North Anna NPS, 25 of 27 spent fuel dry storage casks shifted.

Dry casks for spent fuel moved at North Anna NPS storage area (VEPCO).

Concrete damage: At Fukushima Dai-ichi Unit 6 turbine building, the concrete pedestal of the feed water heater chipped at the edges,

Fukushima Dai-ichi Unit 6 concrete pedestal of feed water heater in the turbine building (TEPCO).

while at North Anna NPS the concrete pedestal of a demineralizer chipped at the edges.

North Anna NPS turbine building chipping of concrete base for demineralizer (VEPCO).

The above similarities suggest that it may be imprudent to underestimate the possibility of damage to North Anna NPS comparable to that observed at nuclear power stations in Japan after a 6-upper quake, because of similar differences between ground motion anticipated in design and ground motion actually occurring. Notably, differences in ground motion between the reactors built on rock and ancillary structures and components built on soil may exacerbate quake impact on their connections.

Outlook
Learning lessons from the Japanese experience, three areas of focus important to safety at North Anna Nuclear Power Station emerge:

Seismic risk reduction through automatic seismic scram
Only few commercial nuclear power stations in the U.S. are outfitted with the equipment needed to automatically scram the reactors when they are subjected to ground motion reaching stipulated set points. The Sep. 8 meeting revealed that the control room operators at North Anna NPS lacked information on containment motion when they needed it most, because at the time of the quake seismic instruments had lost power. The alarms that would have informed the operators whether the ground motion warranted a shutdown failed. On Sep. 28, VEPCO informed the NRC that negative power flux rates at the station caused the scrams. The precise cause of the of the negative flux rates remains to be resolved. Automatic seismic scram systems may have benefitted the operators in shutting down the station's two reactors more expediently. VEPCO should be asked to evaluate whether state-of-the-art automated seismic scram systems would reduce current levels of seismic risk. Moreover, it seems prudent to independently retrofit all seismically important instrumentation with 24 hour-lived batteries. Additional, solar power may extend battery life.
Seismic risk re-evalution of the ultimate heat sink
The reactor core decay heat persisting after a scram must be transferred to the ultimate heat sink. For example, North Anna NPS uses Lake Anna, a dammed lake created for the purpose, as ultimate heat sink in a one-through design. The heat transfer to the ultimate sink is crucial to cooling the reactors, preventing fuel core meltdowns with disastrous consequences as the reactor accidents at Fukushima 1 NPS last March strikingly demonstrate. The preservation of the ultimate heat sink and its use via residual heat removal systems are of existential importance to reactor safety.

Lake Anna's water level had fallen two feet after the Virginia Earthquake Aug. 23, 2011 (source: WTVR CBS 6 Richmond, VA) [17].
According to CBS Channel 6 News, Richmond, VA, the water level of Lake Anna had fallen roughly two feet after the quake [17]. At the Sep. 8 meeting in Rockville, VEPCO stated that an inspection of the dam provided no evidence of quake damage. At the Oct. 21 meeting, VEPCO elaborated that Lake Anna and its main dam do not represent the ultimate heat sink for the station, but that this function is served by a smaller body of water separated from the lake by another dam. VEPCO inspected this dam and found no damage. Regardless of this barrier, the lake represents the ultimate heat sink, if only indirectly. The question what might have precipitated the reported rapid loss of water from the lake remains to be answered. Uncovering the cause for such drastic water level diminution, should it be confirmed, might be of profound importance for the seismic safety of North Anna NPS, because the Lake Anna constitutes the ultimate heat sink for the decay heat of the reactor fuel cores during shutdown.
Seismic risk re-evaluation of all emergency core cooling system SSC
All SSC of the emergency core cooling system, including condensate storage tanks, are important to seismic safety and should be tested structurally and functionally. Particular attention must be paid to SDC-II and III SSC built on soil that are connected with SDC-I SSC built on rock, because soil ground motion is commonly greater than rock ground motion. Differences in ground motion between rock and soil may result in damage to the connections between SSC built on either substrate. At Fukushima Dai-ichi NPS, efforts to prevent the meltdowns of the fuel cores of units 2 and 3 were rendered ineffective by the failure of reactor core isolation cooling (RCIC) systems and high pressure coolant injection (HPCI) systems. The meltdown at Unit 1 was precipitated by a failed isolation condenser (IC). Combined with the failure of the Residual Heat Removal (RHR) system primarily attributed to tsunami damage, these shortcomings contributed in no small part to the failure of the emergency core cooling system (ECCS). The inspections at North Anna NPS must ensure that all SSC of the ECCS conform to the highest seismic safety level and can withstand the ground motion the station is anticipated to incur by substantial margin.

We shall see how VEPCO's re-evaluation of North Anna NPS proceeds. Like at Kashiwasaki-Kariwa NPS Unit 7, function tests of ECCS components during restart will show whether the reactors can be operated safely. Decisive insights will be gained at the very end.

Acknowledgment
I am indebted to the numerous commenters on scribblelive.com Japan Earthquake5 without whose up-to-date input I could not have written this post. The information on www.simplyinfo.org was of invaluable help. I thank MJ Racer whose cogent post on physicsforums (comment number 10945) provided the lead for this essay.

References

Report of Japanese Government to the IAEA Ministerial Conference on Nuclear Safety -The Accident at TEPCO’s Fukushima Nuclear Power Stations- submitted to the International Atomic Energy Agency Jun 7, 2011.
Additional Report of Japanese Government to IAEA - Accident at TEPCO's Fukushima Nuclear Power StationsTransmitted by Nuclear Emergency Response Headquarters, Government of Japan, 15 Sep 2011
Kimura S (Sep 24, 2011) Cesium exceeding safety threshold detected in Fukushima rice. Asahi Shimbun, Tokyo
NHK World News (Jul 19 ,2011) Govt bans beef cattle shipments from Fukushima. Japan Broadcasting Corporation, Tokyo.
The Japan Times Online (Oct 12, 2011) Yokohama finds high strontium-90 levels. The Japan Times, Kyodo.
Kayen R, Collins BD, Abrahamson N, Ashford S, Brandenberg SJ, Cluff L, Dickenson S, Johnson L, Kabeyasawa T, Kawamata Y, Koumoto H, Marubashi N, Pujol S, Steele C, Sun J, Tanaka Y, Tokimatsu K, Tsai B, Yanev P, Yashinsky M, Yousok K (2007) Investigation of the M6.6 Niigata-Chuetsu Oki, Japan, Earthquake of July 16, 2007: U.S. Geological Survey, Open File Report 2007-1365.
Measures taken by the Nuclear and Industrial Safety Agency concerning the Kashiwazaki-Kariwa Nuclear Power Station, affected by the Niigataken Chuetsu-oki Earthquake (2nd Interim Report), June 29, 2009, Nuclear and Industrial Safety Agency.
TEPCO Press Release (Jul 30, 2007) The (First) Report on the Analysis of Observed Seismic Data Collected in Kashiwazaki-Kariwa Nuclear Power Station on the Occasion of the Niigata-Chuetsu-Oki Earthquake in 2007. (Appendix)
Naeim F, Kircher CA (2001) On the damping adjustment factors for earthquake response spectra. Struct Design Tall Build 10: 361–369.
Gaku, S (2010) Experience of NCO Earthquake and Restart of Kashiwazaki-Kariwa NPP. 1^st Kashiwazaki International Symposium on Seismic Safety of Nuclear Installations and Embedded Topical Meetings: C-13.
Antaki G, Johnson J (2011) Seismic design and retrofit of essential systems in nuclear power plants. Becht Nuclear Services.
Ochiai, K, Kobayashi K, Chigama A (2010) Damage Indicating Parameters and Damage Modes of Mechanical Components. 1^st Kashiwazaki International Symposium on Seismic Safety of Nuclear Installations and Embedded Topical Meetings: C-24.
Additional Report of Japanese Government to IAEA - Accident at TEPCO's Fukushima Nuclear Power StationsTransmitted by Nuclear Emergency Response Headquarters, Government of Japan, 15 Sep. 2011.
Leith W (2011) Comments on the Japan Near-Term Task Force Report.
Joe Dwinell J, Sherman N (Aug 23 ,2011) Virginia quake shakes up Boston. The Boston Herald, Boston.
Campbell KW, Bozorgnia Y (2010) Analysis of cumulative absolute velocity (CAV) and JMA instrumental seismic intensity (I_JMA) using the PEER-NGA strong motion database. PEER 102.
CBS 6 WTVR News (Aug. 25, 2011) Mineral residents concerned about Nuclear power plant. WTVR CBS 6 Richmond.

Related Posts

Addendum

In response to the comment to this essay, I provide a Google maps image of North Anna NPS. Looking over the station's lay-out, I find it difficult to conceive that Lake Anna does not ultimately represent the ultimate heat sink for the station as the commenter claims. I leave the decision to the experts (06/12/2011).

Bird's eye view of North Anna NPS. Note Lake Anna and two separate small bodies of water on the premises of the station that could potentially serve as the ultimate heat sink (courtesy Google maps).

The True Speed of Our Net

2011-10-10T09:59:00.000-05:00

Have you ever wondered what internet speed your provider truly delivers. According to this google gadget, the speed in our home is a fraction of the promise. Test it yourself and be surprised!

Kinect on OSX Snow Leopard

2010-11-23T13:02:00.023-06:00

Exhilarating Madness!

Microsoft's XB360 Kinect Sensor reached the market the other day to enrich games on the Microsoft Xbox 360 Console with biological motion detection and superb three-dimensional enhancements. Kinect has received good marks to date. The possibility of creating three-dimensional images with Kinect caught my eye yesterday in a post with the title "With Kinect Controller, Hackers Take Liberties" by Jenna Wortham published online in The New York Times Nov. 21, 2010. The potentials are intriguing. Watch Oliver Kreylos' demonstration below:

I decided to give this a try and compile the software driver needed to run the sensor on a MacBook running Snow Leopard (OSX 10.6.5). Sean Nicholls provides useful information in his post with the title "Compiling OpenKinect on a stubborn OSX 10.6.5 (Updated 21/Nov)", first published Nov. 13, 2010. Sean's instructions are straight forward, leading to a functional application that provides as with a simple start. Below I added a few annotations that may be helpful to the novice builder.

In order to obtain the source codes and compile them into binaries, we need to install the packages 'git' (git-1.7.3.2-intel-leopard.dmg) and 'cmake' (cmake-2.8.3-Darwin-universal.dmg), respectively. Both can be downloaded ready for installation from the links Sean provides in his post. Note that third party unix programs traditionally are installed into the directory /usr/local. The header files needed for compiling other packages are installed in the directory /usr/local/include. Compiled executable binaries will be installed in /usr/local/bin. Libraries will be found in /usr/local/lib, and files with information to be shared in /usr/local/share. This is the default map for the directory tree of all installs that conform to GNU conventions.

However, some software designers prefer to have their packages installed into their own directory tree in a separate subdirectory to /usr/local. This is the case with 'git'. We will find the package in /usr/local/git. If we wish to use 'git' commands in a terminal without providing the full path where they are located each time, we must add this path to our environmental PATH variable just so:

export PATH="$PATH:/usr/local/git/bin

If we now type:

echo $PATH

The terminal should print a colon-separated chain of bin directory paths ending with /usr/local/git/bin.

The next package to install is 'cmake'. The installer will prompt us to provide a directory where to install the executables; the default is /usr/bin. This is a directory for commands that the operating system uses. To avoid mixing system commands with others, I chose /usr/local/bin according to the above conventions. Again we add /usr/local/bin to our PATH variable as suggested above. Then we create a folder for the driver application, change to this directory in the terminal in which we previously set the PATH variable, and follow Sean's further instructions.

Note our additions to the PATH variable pertain only to this terminal window and are lost when we close it. To make them permanent, we need to edit the PATH variable listed in the .bashrc file in our home directory. But this is not strictly necessary for our purpose here.

Sean instructs us to download the sources for two packages that we must compile ourselves, using git: OpenKinect and LibUSB. The first depends on the latter. Hence, LibUSB must be patched for the use with Openkinect, built and installed first, using the OSX generic make command. Note that the package will install into the appropriate subdirectories of /usr/local.

In the next step, we build OpenKinect, more precisely libfreenect, with cmake. The compiler needs to be pointed to the header files that come with LibUSB. This is accomplished with providing /usr/local/include/usblib-1.0/ on the appropriate line in the cmake configuration gui opened using the ccmake command as Sean illustrates. After cmake has configured the build successfully, we compile the package using the make command.

In my configuration, the build also needed to have the path to libjpeg specified, in order to successfully complete its last stage known as linking. The compiler complaint reads:

Linking C executable as3-server

ld: library not found for -lJPEG

In OSX, this library is located in: /System/Library/Frameworks/ApplicationServices.framework/Versions/A/Frameworks/ImageIO.framework/Versions/A/Resources/. There must be ways to configure cmake with this path, probably by adding "-framework ApplicationServices -framework CoreFoundation" to the LDFLAGS variable. Proper configuration is crucial, if we wish to deploy our binaries to other systems. Because I was not planning to deploy the binaries, I decided to avoid re-configuration. Since the compiler always checks /usr/local/lib, I helped myself with providing a soft link in /usr/local/lib to the library's true location just so:

sudo ln -s /System/Library/Frameworks/ApplicationServices.framework/Versions/A/Frameworks/ImageIO.framework/Versions/A/Resources/libJPEG.dylib /usr/local/lib/libJPEG.dylib

With the repeat of the make step the build finished successfully and was ready to install.

Addenda

I just received the Kinect and have been running a test for close to one hour without fail (11/29/10).

Kinect on the Mac

The next step to more functionality entails hooking the kinect video stream up to an Actionscript3 flash server. The openkinect package provides help. Juan Carlos del Valle posted a video of the procedure to be followed with the title "Kinect now accessible from Flash" on the Freenect - Kinect projects site. The video moves along at a fast clip. I had to review the steps over and over again. Below I wrote out a step-by-step protocol of the procedure.

Juan uses two Adobe Flash ActionScript files named test_depth.as and test_rgb.as which can be found in the libfreenect/actionscript directory. We need to download Adobe flex SDK_2.4 from the Adobe flex.org site to accomplish this. Once unzipped, we move the folder to our preferred location. The folder contains a bin-directory that is home to the mxmlc executable we need to use. Hence, it is helpful to add the path to this bin-directory to our PATH variable on the command line in the terminal window, as we added paths to bin-directories earlier. Then, we change into 'libfreenect/actionscript' directory using the cd command and type:
- mxmlc test_depth.as
With a little delay the terminal will respond:
"Loading configuration file {your path}t/flex_sdk_4.1/frameworks/flex-config.xml
{your path}/kinekt/libfreenect/actionscript/test_depth.as: Warning: This compilation unit did not have a factoryClass specified in Frame metadata to load the configured runtime shared libraries. To compile without runtime shared libraries either set the -static-link-runtime-shared-libraries option to true or remove the -runtime-shared-libraries option."

Juan Carlos ignored the message. The command produces the file test_depth.swf in the same directory. We repeat the last step for test_rgb.as to generate test_rgb.swf.

Next we need to start the Actionscript3 flash server. The subdirectory 'libfreenect/c/build/examples' containing the glview executable which generated the twin view in our earlier work is also home to the as3-server excecutable. Because I had not added this directory to my PATH variable, I changed into it and typed:
- ./as3-server
The terminal responded:
Number of devices found: 1
### Wait depth client
### Wait rgb client
### Wait data client

Opening test_rgb in my browser produced a live test image for color beginning with the comment:
### Got rgb client

We kill the server with ctr C, and restart it with:
- ./as3-server
Then we opening test_depth.swf in our browser, producing a live image for depth beginning with the comment:
### Got depth client

However, note that the red channel is not working. I am investigating (12/01/10).

The red channel works with as3-server built using a current libfreenect clone obtained with:
- git clone git://github.com/imekinox/openkinect.git
In order to prompt cmake to include as3-server in the build, we must first change the target option for as3-server in CMakeLists.txt to "ON" (12/07/10).

The Stuxnet Worm, Windows & The Internet

2010-09-26T10:21:00.050-05:00

A malware known as Stuxnet has attracted major media attention in recent days. This software inserts itself into computer systems using Microsoft Windows, Microsoft Explorer, and WinCC developed by the German electrical engineering giant Siemens AG. The latter piece of software is a Supervisory Control and Data Acquisition, or SCADA, application named Simatic S-7 that controls large scale industrial processes like power plants.

Stuxnet is a type of computer worm. That is, the program self-installs executable files that can be hidden on a USB flash memory stick or hard drive to the host computer. Subsequently, the executables are capable of transferring other files through the internet, if a connection is available. Stuxnet checks for narrowly specified system configurations and only acts if WinCC is running. Therefore, its mission consists of taking control of only a small number of exclusive targets.

An East European security firm was first to report Stuxnet last July. The worm is believed to have been active for about a year and has caused considerable disruptions in Asia, notably Iran. Potentially, the program may be capable of disrupting plant operations leading up to the destruction of the facilities. Its release is considered the first discovered cyber attack meant not only to disrupt information technology, but furthermore destroy capabilities.

The Chertoff Group's internet security expert David Falkenrath provides interesting insights on Stuxnet's impact and its ramifications in this interview by Bloomberg's Deidre Bolton entitled "Virus May Target Nuke Plant" aired Sep. 24, 2010.

Bushehr Nuclear Power Plant, Iran (Bloomberg, 2010)

Supposedly the vulnerabilities in Microsoft Explorer exploited by Stuxnet have been plugged. What may be of interest to the common user is that the worm used Microsoft approved security signatures from the network interface card chipset maker Realtek and the flash memory controller developer JMicron to install its files via the internet in the disguise of seemingly legitimate Microsoft-certified driver updates.

We should be safe as long as we keep our system and browser up to date with the latest security patches and avoid legacy hardware with obsolete drivers. That is, we should upgrade to the newest generation of internet adapter cards and regularly update the drivers, downloading directly from manufacturer sites.

Addenda

In 2007, a large-scale bribery scandal broke in Germany, implicating Siemens AG's business in Southeast Asia. By 2008, a number of employees involved in this affair were let go. It only takes one disgruntled software engineer with intricate knowledge of the SCADA program running the targeted facility, maybe with the help of one or two other hackers knowledgeable in Microsoft Explorer and USB driver vulnerabilities, to accomplish Stuxnet in revenge. Perhaps, Iran is a clever diversion, Siemens already paid, and we never find out the actually intended target (10/04/10).
According to William J. Broad and David E. Sanger's article with the title "Worm Was Perfect for Sabotaging Centrifuges" published online in the New York Times today, recent results from the ongoing examination of Stuxnet code suggest that the worm was meant to target the speed control of ultra-centrifuges as those used for uranium enrichment, revving up their speeds to destructive levels. Since the implicated controllers were identified as products manufactured by companies in Finland and Iran, uranium enrichment facilities in Iran may have been the target (11/18/10).
According to William Broad, John Markoff and David Sanger's article with the title "Israel Tests on Worm Called Crucial in Iran Nuclear Delay" published online in The New York Times yesterday, more signs point to Israel and the U.S. as Stuxnet's originators with centrifuges in Iran's Natanz uranium enrichment plant as the target (01/16/11).
John Markoff reports in his article with the title "Malware Aimed at Iran Hit Five Sites, Report Says" posted online Feb. 11, 2011, that according to a Symantec study Stuxnet may have infiltrated as many as five Iranian institutions in three, possibly four, waves (01/13/11).
Listen to KCRW TO THE POINT's Warren Olney interview David Albright, President of the Institute for Science and International Security, on Stuxnet's effects in Iran 44 minutes into today's show with the title "A New Paradigm in the Middle East". According to Albright, Stuxnet disrupted the Iraninan uranium enrichment program noticeably, but only in small ways. The greater threat may lie in the potential of adversaries using the worm, now public, as prototype for future attacks elsewhere (02/16/11).
According to Noah Shachtman's WIRED report with the title "Computer virus hits U.S. drone fleet" published online on CNN today, U.S. drones have been infected with a tracking virus, possibly through USB flash memory devices. This is not quite stuxnet yet, but a first step seems taken (10/10/11).

Related Posts

Windows 7, Qemu & the Internet

Sources

ako (07/28/08) Siemens-Prozess: Erstes Urteil im Schmiergeld-Skandal. Deutsche Welle.
Gjelten T (09/27/2010) Who's Behind Cyber-Worm Targeting Iran? National Public Radio's All Things Considered.
Gorman S (/09/27/10) Computer Worm Hits Iran Power Plant. The Wall Street Journal.
Gorman S, Fidler S (09/25/10) Cyber Attacks Test Pentagon, Allies and Foes. The Wall Street Journal.
Li P (08/22/07) Siemens China Linked to Bribery Scandal. China.org.cn.
Maclean W (09/24/10) Cyber attack appears to target Iran: tech firms. Reuters.
Richmond R (09/24/10) Malware Hits Computerized Industrial Equipment. The New York Times.
Sanger, DE (09/25/10) Iran Fights Malware Attacking Computers. The New York Times.
S.T. (09/24/10) A cyber-missile aimed at Iran? The Economist.

Solar Power: Way to Go!

2010-09-15T09:21:00.009-05:00

We own a 12.0-inch diameter, 9.0-inch high urn-shaped fountain carved from limestone by the Tennessee artist Scott Wise. The fountain features a round shallow indentation on top into which water needs to be pumped slowly through a 0.5-inch center bore to feed a gentle, steady glistening pour over the ledge and down the sides of the rugged surface of the stone.

I set the fountain on two weathered brownstones in a 14.5-inch diameter, 4.0-inch deep pan sold for changing automotive oil, threaded a plastic hose through the center bore, sealed it with Macco Adhesives LN903 Liquid Nails Heavy-Duty Construction and Remodeling Adhesive, and connected it to a small pump nestled between the stones. I balanced the fountain stone with small flat river stones to allow the water flow down evenly on all sides. The pan holds more than enough water for the fountain to perform properly. However, owing to evaporation we need to top the water in the pan at least once a week to protect the pump from overheating. If we live in a dry climate, we need to check more regularly. At our first home, I used a pump fed by a power line laid underground in 0.75-inch diameter PVC tubing. I had to dig a foot-deep trench from our home across the entire yard. The job meant considerable work for a small flow of water.

To minimize the effort at the fountain's new location, I opted for a solar-powered pump which I purchased from Shore Power Inc. / BatteryJunction.com.

Power is generated with a 5.875-inch by 9.625-inch solar panel that can be installed at a distance from the fountain. The pump is equipped with a 14-foot long power cable. Where we live in the Southeastern U.S., the setup works great. However, the panel's effective angle is narrow; it must face the sun directly to produce the greatest power. We had to place it in the sunniest spot of our yard. Perhaps, mounting the panel on top of a wind-up clockwork that turns it along the sun's course is worth a consideration. Regardless, the smallest cloud blocking the sun will shut the fountain down. Despite, I got used to the intermissions, enjoying the sight of the bubbling little flow glistening in the sunlight as an affirmation that solar power does work indeed. However, if we wish to power our home with solar panels, it must be located in a particularly sunny spot, we probably need plenty panel surface, and the panels must be oriented such that a sufficient number face the sun at any time of the day. My small experiment is supposed to constitute an affordable test, informing us on the costs we may have to anticipate, if we aspire to convert our home.

Addendum

Solar power windows sound like a great idea. Watch this Reuters report (08/11/2011):

Nexus One: The Google Phone

2010-01-06T10:07:00.008-06:00

With Google's Nexus One officially introduced to the world yesterday, a new potent competitor has entered the highly-contested smartphone market. I won't comment much on the phone's features. Nexus One appears to be a solid product based on Google's Android operating system and Google applications not unlike the Droid introduced late last year. However, while the Droid is only available locked in with a particular cell phone service provider, the customer will be free to choose a plan to her/his liking with the Nexus One.

By contrast to their strongest competitor, Apple's iPhone, Android-based smartphones are open to Java programming. This constitutes a great advantage for application developers. However, beware! The libraries are restricted. Conventional Java applications will not run without major tweaking.

The footage below highlights last week's news development on the issue:

Review (01/05/10)

Revelation (01/05/10)

Rumor (12/31/09)

Addenda

Alexei Oreskovic reported in his post entitled "Google's Android takes No 2 spot from iPhone" on Reuters yesterday that according to NPD Group market research Google's Android has become at 28 percent market share the second most prevalent smartphone operating system in the U.S. after Symbian used on Blackberry (36 percent market share) in this year's first quarter, displacing Apple's IPhone using OS X (21 percent market share). The figure comprises all phones running Android. Google's own Nexus One has not had such great success yet (05/11/10).
According to Doug Gross' CNN post today entitled "Google quietly kills its once-hyped Nexus One phone", Google Inc. decided to stop selling its own smartphone Nexus One entirely because of sluggish sales. Nexus One uses Google's Android operating system. Other smartphones with the same operating system like HTC's Droid Incredible are selling briskly, and Google will maintain full support for the operating system (07/19/10).

A First Peek at Google's Chrome OS

2009-11-25T16:18:00.013-06:00

Last Thursday, Google Inc. introduced Chrome OS aka Chromium OS, to the public with quite considerable fanfare. The news enjoyed universal coverage in the media. Chromium OS is an operating system that the company is developing on its own as an open source effort. Sources and build directions are now easily accessible.

I gave it try. In short, Chromium OS consists of a Debian-derived operating system using a Linux 2.6 kernel and Google's Chrome browser. It is meant to run on personal computers with the i386 architecture with either 32- or 64-bit processors. With the packages Google includes, the compressed install image is roughly 800 megabytes in size and takes up 2.8 gigabytes of disk space expanded.

The distribution is open to customization. Adding applications is possible. Only, the packages have to comply with the Debian distribution format.

Obviously, Chrome OS is meant to run bare-bones on very thin clients. Non-essential applications and data are intended to be stored elsewhere in the cloud. The small footprint and, hopefully, the ensuing speed may harbor this operating system's greatest strengths, potentially turning Chrome OS into a powerful, cost-efficient tool for corporate users.

In view of the current governmental push for patient record digitization, the health care sector appears particularly suited. I regularly visit a large academic outpatient clinic. Every examination room is equipped with a conventional stand-alone personal computer connected to an intranet. Across the entire hospital campus, these computers must number in the thousands. In such environment, thin clients appear superbly suited to drastically diminish cost for hardware, software and IT maintenance. Here, Chrome OS may provide a rich resource.

The build instructions for Chrome OS on the Chromium project site are straight forward. I compiled the system with Ubuntu's Karmic Koala (9.10) hosted on a work station with two 64-bit AMD Opteron 246 processors. The hardware is described in detail in my post dated May 22, 2008. I opted not to compile the Chrome browser myself. Instead, I downloaded the pre-built package available from Google and copied it to the build directory as directed in the Chromium project documentation. Do not omit to provide a user name and a shared user password. They may be essential for log in and system administration later.

The build process unfolded uneventful. However, complications arose, when I had to decide how to actually run the Chrome OS.

Google provides two options to ready the successful build for installation. One method uses a bootable USB 1-gigabyte flash memory drive, from which the system can be installed on a dedicated computer. For this procedure to work, USB devices have to be enabled in the boot list of the bios first.

The other installation option bundles the build in a VMware image named ide.vmdk.

Since I did not wish to dedicate a computer to testing Chrome OS quite yet, I chose this option. Eventually, I managed to install the image on a virtual personal computer emulated with qemu-kvm (version 0.11.0). Unfortunately, the emulation slows operations down considerably. Regardless, the procedure worked sufficiently fast for preliminary exploration.

I used qemu with the following options on the command line to start the emulator:

qemu-system-x86_64 -localtime -m 256 -vga std ide.vmdk

The screenshots below illustrate the sequence of steps I encountered starting up Chrome OS.

After a successful boot, the system introduces itself with a login screen.

I used my preset shared user password to log in. After we are signed in successfully, the Chrome browser launches. automatically.

We may browse the web or start our web applications.

Not too bad at this point!

Alas, I was not able to bring Chrome OS up to speed with qemu. I found a satisfactory solution in virtualbox. This emulator, however, does not accept the format of ide.vmdk. It has to be converted to vdi. The conversion takes two steps: We have to convert the Chrome OS vmdk image to a raw image with the suffix .bin with two commands in the directory where the image resides:

We have toconvert the Chrome OS vmdk image to a raw image with the suffix .bin using qemu-img convert ide.vmdk chrome_os.bin
Then, we convert the raw image into a vdi image using a virtualbox tool with the command VBoxManage convertdd chrome_os.bin chrome_os.vdi

If we have not installed virtualbox yet, we shall be instructed to do so using sudo apt-get while trying to execute the second command.

After opening the virtualbox graphic user interface with the command virtualbox, we attach the vdi image as a hard drive from the menu by clicking Add in the Hard Disks tab under File > Virtual Media Manager. We locate and select the chrome_os.vdi image and click Open. Once the image has been added to the list, we click OK, and continue to create the remaining virtual machine profile. I reserved 256 megabytes as Base Memory. The operating system is Linux. The version is 2.6. The rest is set by default or optional. Now, Chrome OS is ready for boot.

Virtualbox is also available for Apple Macintosh computers with the intel architecture. Following the same steps as above, I successfully installed Chrome OS using the vdi image in virtualbox on a Mac Mini running OS X 10.5 (Leopard).

The emulator runs Chrome OS on both computers surprisingly fast.

Addendum

According to Stephen Shankland' s post with the title "Google shows off Chrome OS tablet ideas" on CNET.com dated Dec. 29, 2009, that I found on CNN.com today, the ideal thin client running Chromium could be a tablet (02/07/10).

Related Posts

Build your own PC: Episode I

Related Links

Alexei Oreskovic's report on Reuters entitled "Google shows off Chrome operating system" and dated Nov. 19, 2009.
Miguel Helft's article in The New York Times entitled "Google Offers Peek at Operating System, a Potential Challenge to Windows" and dated Nov. 19, 2009.
Jessica Vascellaro and Scott Morrison's article in The Wall Street Journal entitled "Google Outlines Chrome OS Plans" and dated Nov. 19, 2009.
David Goldman's post on CNN.Money.com with the title "Tablet wars: Google looks to take on Apple iPad" dated Feb. 4, 2010.

Chinese Drywall: kleiner Hinweis!

2009-11-23T11:32:00.014-06:00

Auf Grund des rasanten Immobilienpreisverfalls in den amerikanischen Golfanrainerstaaten und des zur Zeit niedrigen Dollarkurses bereisen wachsende Zahlen deutscher Interessenten Florida auf der Suche nach günstigen Ferienbehausungen.

Sollten Sie den Ankauf einer Immobilie in Florida ernsthaft in Erwägung ziehen, achten Sie darauf, daß bei Reparaturarbeiten keine Rigipsplatten aus China verwendet worden sind.

Die tropischen Stürme der letzten Jahre haben an vielen Gebäuden an der Golfküste beträchtliche Wasserschäden hinterlassen. Der enorme Bedarf an Baumaterial hatte die Liefermöglichkeiten heimischer Hersteller schnell erschöpft. Importeure, zum Teil aus Deutschland, sprangen mit Billigware ein. Vielfach wurden zur Wandinnenverkleidung Kartongipsplatten aus China eingebaut.

Es stellte sich schnell heraus, daß diese Rigipsplatten stark korrosive Stoffe (vor allem Schwefelwasserstoff und Formaldehyd) in hoher Konzentration enthalten. Dies führte in kurzer Zeit zu erheblichen Schäden an elektrischer und sanitärer Installation. Darüber hinaus klagen einige Bewohner über chronische Gesundheitsbeschwerden.

Die Rigipsplatten müssen in diesen Häusern letztendlich ausgetauscht werden. Die geschädigte Infrastruktur, inklusive Heizung und Klimaanlage, muß ersetzt werden. Die Umbauarbeiten sind mit erheblichen Kosten verbunden, die keine Versicherung erstattet.

Bestehen Sie deshalb vor dem Kauf ihres Ferientraumes im Sonnenscheinstaat auf eine rechtsgültiges Gutachten, daß bei Instandsetzungsarbeiten keine chinesischen Rigipsplatten verwendet worden sind. Falls dies der Fall gewesen sein sollte, müssen die Schäden adequat behoben sein. Gutachten und Schadstofftests sind vor Ort erhältlich. Immobilien ohne Zertifikat sind weder versicherbar, noch weiterverkäuflich!

Anhang

Immobilienversicherungen sind in Florida recht teuer, insbesondere an der Küste. Flut- und Wasserschäden sind in Minimalversicherungspolicen nicht inbegriffen. Die staatliche Zusatzversicherung deckt Windschäden ab. Es ist darauf wertzulegen, sich neben Feuerschäden auch gegen Flut- und Wasserschäden abzusichern. Schäden durch starken Regen und Flutwasser treten in Florida häufig auf. Mit ihnen muß gerade bei tropischen Stürmen gerechnet werden. Je nach Lage und Wert der Immobilie kann eine Versicherungsprämie $15.000. im Jahr überschreiten.
Bob Van Voris and Allen Johnson Jr. berichteten letzte Woche in ihrem Bloomberg News Artikel mit dem Titel "Judge Awards $2.6 Million in Chinese Drywall Suit", daß ein Bundesrichter in New Orleans im ersten Urteil über eine Klage von geschädigten Hausbesitzern gegen den Hersteller chinesischer Kartongipsplatten im Sinne der Kläger entschieden hat. Das Urteil wird einer Vielzahl ähnlicher Klagen als Vorbild dienen (13.4.10).
Zu Folge des heutigen Berichts von Robbie Whelan and Dawn Wotapka im Wall Street Journal mit der Überschrift "Wallboard Exporter in Settlement With Beazer Over Costly Repairs" ist einer der größten Vertreiber chinesischer Kartongipsplatten in den USA, Knauf Plasterboard Tianjin Co mit Sitz in Deutschland unter dem Namen Knauf Gips KG, bemüht, durch zügigen Vergleich teure Gerichtsprozesse mit geschädigten Baufirmen zu vermeiden (16.5.10).
National Public Radio berichtete heute in seinen Nachrichten unter dem Titel "Jury Awards $2.4M In First Chinese Drywall Trial", daß ein Jurygericht in Florida einem Ehepaar, dessen Neubau durch chinesische Gipsplatten ruiniert wurde, Schadenersatz in der Höhe von 2,4 Millionen Dollar zugesprochen hat. Der Betrag ist zu 55 Prozent vom Vertreiber der Platten, Banner Supply Co., zu bestreiten. Der Rest soll vom Hersteller Knauf Plasterboard Tianjian getragen werden (19.9.10).
M.P. McQueen berichtet in seinem Betrag mit der Überschrift "Chinese Drywall Settlement Reached" im Wall Street Journal, dass sich heute im chinesischen Kartongipsplatten Prozess in New Orleans die klagenden Eigentümer von 300 Häusern in vier Staaten mit dem Plattenhersteller Knauf Plasterboard Tianjin (KPT) auf einen Vergleich geeinigt haben. KPT, die Zulieferer und deren Versicherungen erklären sich bereit, die Kosten für den Austausch der Platten und alle notwendigen Hausreparaturen zu tragen, Klimaanlage ausgeschlossen. Die Beklagten gestehen dabei nicht ein, dass die von den Klägern aufgeführten Schäden auf die fehlerhaften Platten zurückzufüheren sind. Die Reparaturarbeiten sind von den Beklagten bewilligten Bauunternehmen durchzuführen. Kosten für Ersatzwohnungen während der Reparaturarbeiten und die damit verbundenen Umzüge werden ebenfalls bestritten. Es ist anzunehmen, dass dieser Vergleich noch ausstehenden Klagen zur Vorlage dienen wird. Klagen gegen das deutsche Unternehmen Knauf Gips KG sind noch nicht entschieden. Die Firma steht auf dem Standpunkt, mit KPT nur eine lose Partnerschaft ohne jegliche Kontrolle zu unterhalten und daher nicht für das Geschäftsgebaren von KPT verantwortlich zu sein (14.10.10).
Nach Dawn Wotapkas Artikel mit der Überschrift "Homeowners to Be Repaid in Drywall Settlement", der heute im Wall Street Journal erschien, könnte es in der 'class-action' Schadenersatzklage gegen KPT und anderen chinesischen Kartongipsplattenherstellern im frühen nächsten Jahr zu einer endgültigen Einigung kommen (15.12.2011).

Lesestoff zum Thema:

Beitrag von Charlotte Cuthbertson, erschienen am 17.4.2009 in der Epochtimes, mit der Überschrift " Giftige chinesische Gipsplatten von Knauf in den USA aufgetaucht."
Beitrag von Michael Corkery, erschienen am 17.4.2009 im Wall Street Journal, mit der Überschrift "Homeowner Problems With Chinese-Made Drywall Spread."
CNN Betrag von Rich Phillips vom 23.11.2009 mit der Überschrift "Preliminary reports link Chinese drywall, corrosion in U.S. homes."
Beitrag von Andrew Martin vom 17.09.2010 in der New York Times mit der Überschrift "Drywall Flaws: Owners Gain Limited Relief."
Beitrag von M.P. McQueen, erschienen am 14.10.2010 im Wall Street Journal, mit der Überschrift "Chinese Drywall Settlement Reached".

Netgear WG311v2 Wireless Adapter & 64-bit Operating Systems

2009-09-25T12:41:00.018-05:00

I was trying to make a Netgear WG311v2 wireless PCI adapter connect to our Apple MB763LL/A AirPort Extreme Dual-band Base Station router using WPA password protection. I am running Ubuntu 9.04 (Jaunty Jackalope) on a machine with two AMD Opteron 64-bit CPUs. Note the card works out of the box with the driver provided by ubuntu on unprotected networks as well as on networks with WEP password protection, but not with WPA password protection.

I spent a whole week on the idea of adding WPA capability with the result that VERSION 2 of the adapter JUST WILL NOT work with this option on machines with 64-bit processors, because nobody compiled drivers for the 64-bit architecture. This simple truth applies to ALL wireless adapters based on the TEXAS INSTRUMENTS ACX100/111 aka TNETW1450 chipsets, regardless of brand, and also applies to ALL Windows 64-bit operating system users.

The careful examiner will already notice this bitter truth on the adapter label shown on the Netgear support site. On the label of WG311v2, you will find 32-bit PCI. They mean it! By contrast, WG311v3 uses a Ralink chipset, for which 64-bit drivers are available.

Ubuntu 9.04 supports WG311v2 cards on the AMD 64-bit architecture, using the drivers that the Acx100 sourceforge project provides. This project develops Linux drivers for the acx100/111 chipset without help from manufacturers. In my situation, this meant that the card was immediately recognized after boot and the driver was loaded. I could configure the default wlan0 connection with a WPA key in NetworkManager (0.7.0.100). The spinning wheel in place of the network icon on the menu bar indicated repeated attempts to associate with the router. However, the connection was never successfully established.

I reckoned that perhaps NetworkManager did not set up the WPA password properly and delved into the use of WPA_supplicant for setting up a connection without using NetworkManager. This ubuntu geek tutorial was instructive. Still I could not associate the card with the router.

I decided that perhaps the driver module provided with the ubuntu distribution did not support WPA. I began to search for alternatives. Since Netgear claimed on their site that WPA was supported by their latest driver software (version 2.0.0.7), perhaps there was a way to use their driver.

I happened upon Ndiswrapper. This application wraps windows wireless adapter drivers into a Linux module. Because I could use a wired internet connection temporarily, I was able to install version 1.9 with ubuntu's Synaptic Package Manager. I found good instructions how to proceed on this ubuntu forum here. Most importantly, in order to load the new ndiswrapped acx module successfully, you need to blacklist the old module provided by ubuntu in /etc/modprobe.d/blacklist.conf. The name of the module to be blacklisted is "acx".

Ndiswrapper recognized the card correctly and installed the Netgear driver without complaint. However, the command sudo modprobe ndiswrapper failed.

Why?

Because the Netgear driver was not compiled for a 64-bit architecture.

Remember the label on the card says "32-bit"! The constraint does not stem from the hardware, but from the unavailability of a compatible driver. I checked with all other adapter manufactures using this chipset. No luck!

So unless the next version of ubuntu's acx driver will support WPA, consider buying a new adapter card, making sure that the manufacturer provides drivers for a 64-bit architecture. Since most adapters are available in several versions, make sure to pick the correct one. Remember the difference between Netgear's WG311v2 and WG311v3, I noted above? You may have to check the label on the pci card. Once you obtained an adapter with a 64-bit driver, you may use ndiswrapper, if ubuntu does not support the options you want out of the box.

If I am not mistaken, 64-bit Microsoft Windows users (be it XP_64, vista or windows 7) will run into this incompatibility-of-driver problem as well.

Addendum

The Sabrent PCI-G802 adapter works on Ubuntu 9.10 out of the box. Mine is based on the Ralink RT2561 chipset. If needed, the latest Windows 64-bit drivers are available for download on Ralink's software support site (01/14/10):

National Security & Intellectual Property II

2009-07-09T22:30:00.026-05:00

I have written about the recent purported computer software code theft at Goldman Sachs Group Inc in my post dated Jul. 6, 2009. Developments in the past 72 hours support the idea that Sergey Aleynikov intended to use the computer files he allegedly stole from his former employer in his new job.

He uploaded compressed program code files for an automated low latency stock trading software platform he helped develop at Goldman Sachs onto a server overseas. The number of files was so large that the compressed RAR file archive supposedly topped at 32 MB in size. Assuming text files and 23% compression, Sergey copied roughly a whopping 140 MB of code, that is almost three times the size needed to install a small operating system like Minix. Compiled, the code could produce binaries for a 300 MB application, approximately a third of Microsoft's Office suite.

Today, Svea Herbst-Bayliss and Christian Plumb report in their post on Reuters that the three founding members of the startup Teza Technologies LLC, that had wooed Sergey from Goldman Sachs for three times the salary, were sued by their own old employer, Citadel Investment Group, for violation of noncompete clauses. Citadel of Chicago manages hedge funds. The three worked in the quantitative trading branch of Citadel using automated high frequency methods which yielded extraordinary returns last year.

In all likelihood, the foursome planned to use the fruits of their labor at their respective former employers for the development of an own low latency quantitative trading platform. Could they have manipulated the market? Did their plan threaten national security?

Related Posts

National Security & Intellectual Property

Addenda

According to Laurence Fletcher's post on Reuters this afternoon, Sergey's transferred files were accessible on the server abroad until last Monday. Sergey allegedly transferred them in the days before he left Goldman Sachs at the beginning of June. Hence, there was potential access to the data for about a month (07/10/09).
The owner of the service that hosted Sergey's upload recounts his experience in the past week here. Note that this company's home is in the U.K., whereas its servers are located in Bavaria (07/13/09).
Rob Iati summarized the fundamental implications of low latency high frequency trading (HFT) in his post on advancedtrading.com dated Jul. 10, 2009 (07/17/09).
Today, Jonathan Spicer reports in his post on Reuters about the role of flash programs in high frequency trading, like Direct Edge's Enhanced Liquidity Provider (ELP). Direct Edge flashes stock orders to select costumers for milliseconds, giving them a peak preview of the order flow. In combination with high frequency trading platforms that access dark pools matching orders anonymously, the ELPs may create a two-tired trading system, disadvantaging traders who cannot execute such fast trades. Direct Edge is owned by a consortium consisting of Citadel Investment Group, Goldman Sachs Group Inc, JPMorgan Chase & Co, Knight Capital Group, and International Securities Exchange LLC (07/27/09)
According to Grant McCool's post on Reuters today, the case USA versus Aleynikov (09-mj-01553, U.S. District Court for the Southern District of New York) may find a quiet, anticlimactic ending. In any case, Sergey's alleged theft brought the potential dangers of HFT to the attention of the media and the public (08/03/09).
According to Alex Berenson's post on The New York Times yesterday, the size of Goldman Sachs' HFT program is 1,224 MB. Sergey transferred no more than a quarter of that (08/24/09).
Today, Sergey Aleynikov appeared in federal court for the first time and pleaded "not guilty" to all charges of code theft from Goldman Sachs. The trial is scheduled to begin Nov. 29, 2010. The case is now assigned as "USA versus Aleynikov, U.S. District Court for the Southern District of New York, No. 10-00096" (02/17/10).
According to Chad Bray's report with the title "Ex-Goldman Programmer Found Guilty in Code Theft" published online in The Wall Street Journal today, a jury found Sergey Aleynikov guilty of theft of trade secrets and transportation of stolen property. Read Ted Thomas' insightful comment on high frequency trading added to this report (12/10/10).
According to Grant McCool's post with the title "Ex-Goldman programmer gets 8 years for code theft" published online on Reuters yesterday, Sergey was sentenced to eight years in prison. I presume he will appeal the sentence (03/19/11).

National Security & Intellectual Property

2009-07-07T16:31:00.014-05:00

I am rarely willing to accept conspiracy theories, except maybe the assassination of President Kennedy only because it is highly improbable that a man like Lee Harvey Oswald with the rifle he used could hit a person in a moving car from such distance. Therefore, I do believe in Lee Harvey Oswald's claim that he was "just a patsy" and reckon that Jim Garrison was onto something real after reading his book entitled "On the Trail of the Assassins: My Investigation and Prosecution of the Murder of President Kennedy."Unfortunately, his trails ran cold. We may never find out the truth.

Two days ago another story broke with a whiff of conspiracy. Compared to President Kennedy's assassination, the event seems minor, but appears to fascinate many in the business of stock trades. The incidence was first reported by Tyler Durden in his post on Zero Hedge and Mathew Goldstein in a post on Reuters.

Sergey Aleynikov worked as a programmer in a supervisory function for Goldman Sachs Group Inc. on an application that allows the firm to execute stock trades within milliseconds. The application is known as low latency or high frequency trading (HFT) platform. About a month ago, he started a new post with a start up company for three-times his salary at Goldman Sachs. Before he left Goldman Sachs, he transferred a large volume of files with program code from his computer at the firm to a server in Germany. He encrypted the data and attempted to obscure the transfer.

Weeks later IT security at the firm found out about Sergey's actions and reported the purported "theft" to the FBI. Sergey was arrested last Friday and was freed on $750,000.- bail yesterday, according to Martha Graybow's report on Reuters today. Furthermore, Brent McCool reports on Reuters that U.S. prosecutor Joseph Facciponte told the court in Saturday's hearing of Aleynikov's case that the program in the hands of competitors could cost Goldman Sachs millions. The firm's bearing appears to lend extraordinary importance to what could be just a blatant case of a programmer sloppiness, leaving ample room for speculation about the company secrets that may be contained in the "stolen" files. Conspiracy theories begin to blossom.

Sergey claims in his defense that he intended to copy only opensource files free for anyone's use, but may have included proprietory files inadvertently. He routinely downloaded such files to work on them at home. He did not believe that his actions breached his contract.

Goldman Sachs is a leader in the business of computed millisecond stock trades, reaping great profits from the transactions. The business is highly controversial because of its considerable impact on the market and its potential of market manipulation. When huge volumes of stock are moved in such short time, fortunes are made and lost before anyone without such fast access can respond. The market changes faster than the trader on the floor can pick up his phone. Critics loath Goldman Sachs for their advantage and surmise that the firm is manipulating the entire national economy in its favor with its fast trades. Hence, some hope that this incident may uncover information that precipitates Goldman Sach's demise. Others elevate Sergey's actions to a case of industrial espionage that may endanger national security, since he transferred sensitive information abroad. Hence, Sergey's purported crime quickly garnered intense media attention in recent days.

I see two possible explanations for Sergey's actions:

either Sergey truly did not understand what he was doing when he transferred the data,
or he knew exactly what he was doing, but acted that way in order to be able to claim that he did not know what he was doing,once the data transfer was discovered.

Finding out whether he shared the data with third parties and with whom may provide an answer. The download history should be retrievable from the server in Germany he uploaded the files to. Moreover, anyone who was going to use the files needed Sergey's encryption key.

Furthermore, I reckon that even if third parties got hold of the files, the data would be of limited use to them. Direct implementation is impossible, since the program is dependent on file libraries stored in company-localized systems and needs access to company-specific data bases. In addition, you would have to be situated physically close to the New York Stock Exchange in order to achieve the necessary velocity in data transfer.

Data is transferred across the internet via nodes. The fewer nodes the data has to travel, the faster the transfer. Physical distance adds nodes. Even if a firm in Europe could fully implement Goldman Sachs' program, they would not be able to beat the firm's trades because of the difference in the number of nodes.

On the other hand, if you were installed on Wall Street, close emulation of Goldman Sachs' trading program would be discovered swiftly, and the perpetrator would have to face costly litigation over patent infringements.

Perhaps the third party could examine the strategies and methods used in the programs in an attempt to develop superior ones. I was told that this was an undertaking bound to fail because of the sheer endless lines of code that have to be studied closely and the enormous complexity of such program. Perhaps, savants could run tests on the program to discover vulnerabilities that own programs could exploit to edge out the Goldman Sachs trades. Perhaps, they could uncover the secrets of the decision making engine. I assume you need considerable expertise in the field of fast trade programs in order to accomplish these goals.

The above limitations leave one other possibility that has been suggested as the most likely scenario. Sergey copied the code files for the programs in whose development he was most intimately involved to use them as reference in future work on his new job.

Hence, Sergey's "crime" may consist of nothing more than a misinterpretation of clauses in his job contract and does not merit the media attention it attracted. I find it astounding that a firm like Goldman Sachs was not taking more precautions against such data misuse by employees, particularly when the employees are known to move on to a potential competitor. You only have to monitor the users' shell history and the syslog files. After all, the possibility remains that company secrets vital to the firm's mode of operation are included in the files Sergey knowingly or inadvertently transferred.

We shall find out more. Stay tuned.

Related Posts

National Security & Intellectual Property II

Addendum

On a funny note, according to Nick Carey's post on Reuter's yesterday a business school professor concluded that Sergey's brain told him to transfer the files. Surprised (07/09/09)?

Acknowledgment
I am grateful for the insights of the coding experts posting comments on Zero Hedge. They helped me better understand the issues involved.

Negative Equity, Recovery & Mobility

2009-05-13T16:02:00.034-05:00

Les Christie reported in his post for CNNMoney dated May 6, 2009, that according to a market analysis published on Zillow.com about 22 percent of American homes were "underwater" at the end of March. That is, in today's market the value of these homes is less than the outstanding balance on their mortgages. Home owners may wonder, how long recovery may take. The calculator below provides an estimate, regardless of the price of the home. The default entries are explained in the next paragraph. You may replace them with your own data:

Suppose we bought a home with a down payment of 20 percent and a 30-year mortgage at 5.92% APR two years ago. According to a report on RealEstateabc.com, the annual rate of appreciation for American homes has been on average 6.34 % over the past 40 years. This value seems high. Local rates may be lower. According to my own observations over the past 10 years, and these were years of unprecedented economic growth where I live, property values have appreciated at 5.7% annually. The rate may be substantially less in current circumstances.

Be it as it may, let us take the optimistic view and enter 6.34% as annual appreciation rate in the calculator as default. If we lived in an area of the country hardest hit by the recent slump in the real estate market, our new home may have lost half of its value since we bought it. That is, its value depreciated 50%. According to the result calculated above, we need to keep this home for 21 years to recoup our loss. As a consequence, mobility loses its luster. Many home owners may choose to stay put.

Addenda

According to The Economist's daily chart post dated Aug. 21, 2009, Deutsche Bank's securitization team estimates that roughly every second American home with a mortgage will be underwater in 2011 (08/29/09).
Lisa Lambert reports in her post on Reuters dated Sep. 17, 2009, that another wave of troubled mortgages is about to ensue, potentially unleashing more foreclosures. That is, adjustable rate mortgages with payment options are beginning to reset in large numbers burdening home owners with ever higher cost. In Arizona alone, 128,000 payment option ARMs will adjust to higher rates within the next 12 months. In the meantime, the unemployment rate rose to 9.7 percent nationwide according to the Bureau of Labor Statistics (09/19/09).
According to Les Christie's report entitled "Nearly 25% of all mortgages are underwater" on CNNMoney yesterday, First American CoreLogic estimated that home values are still dropping under the value of their mortage. The company found that in the last quarter of 2009 mortgages on homes underwater increased by one percent to 10.7 million (02/24/10).
M.P. McQueen's post with the title "The New Rules of Remodeling" for the Wall Street Journal dated Apr. 24, 2010, confirms my prediction on rising immobility. Remodeling seems a good idea. Let us benefit from tax credits and put some energy efficient insulation in our home. Let us upgrade kitchen appliances and make our home a nicer place to live, because we may stay in it for a long time (04/28/10).
While modifying my calculator after one reader's cogent comment, I noticed that in the parlance of the Goldman Sachs bankers who testified before Congress two days ago our home, looked at as an investment, definitely qualifies as a very long sale. The banks, however, sold our mortgage short. Our mortgage was sold to Citigroup two years after we closed on the home. Since we could keep up with our payments, it probably ended up buried in tranche A of one of those sh***y CDO's alluded to in the hearing (04/29/10).

According to Conor Dougherty's report entitled "More Americans Moved in '09, but Not Far" in today's Wall Street Journal, the Brookings Institution estimates a state-to-state mover rate of 1.6 percent for 2008 and 2009, constituting the steepest decline in interstate migration since the Great Depression (05/10/10).
You may wish to listen to some homeowners whose diminished home value considerably affected their mobility in Yuki Noguchi's report on National Public Radio's Morning Edition today entitled "Devalued Homes Anchor Prospective Job Seekers" (08/26/10).

Related Posts

Subprime Lending: Truthiness & Delusion of Mind

Windows 7, Qemu & the Internet

2009-04-07T10:18:00.040-05:00

I decided to participate in Microsoft's testing program for Windows 7 earlier this year. I downloaded the 32- and the 64- bit installation package iso-images from Microsoft's website. I envisioned to test the next generation Windows operating system as guest in a qemu emulator (version 0.9.1) using ubuntu's Hardy Heron 8.04 as the host operating system. Hardy Heron is installed on a MSI K8T Master 2FAR motherboard with two AMD Opteron 242 CPUs and 4Gb of RAM. Binary versions of qemu compiled as applications for for Apple's OS X are available from Q. The installation of Windows 7 should not be different. Below I describe what needed to be done to render a successful installation.

KQEMU INSTALL
To install kqemu on the host, I essentially followed instructions provided on the ubuntu community documentation site here and on alien.slackbook.org here. Microsoft recommends to provide at least 16 Gb for Windows 7. I decided to create a raw image of 20 Gb and reserve 1024 Mb for RAM, that is the maximum qemu allows, in the /home directory for the guest, running the following command in the terminal:

qemu-img create -f raw windows7.img 20G

WINDOWS 7 INSTALL
I used the following command to install Windows 7 on the raw image:

qemu-system-x86_64 -localtime -net nic,model=ne2k_pci,vlan=0 -net user -m1024 -cdrom './windows7.iso' -boot d windows7.img

I repeatedly attempted to use the 64-bit version of Windows 7 without success. Each time, the installation process halted with a blue screen below.

By contrast, I was successful with the 32-bit version. The install went smoothly. I could login and test the applications. However, I could not connect to the internet. The internet icon in the bottom panel was crossed out. The problem solver suggested that the proper driver was not found. I was asked to provide one. The reader is advised to review the experience I recount below to the end.

INTERNET
I examined numerous threads on drivers for windows and qemu over several weeks. By default, Qemu provides an ethernet interface that uses the Realtek RTL8029 driver. This is a legacy driver that Realtek no longer supports. I found an installable driver v5.08 here.

Installing, the RTL8029 driver resulted in limited connectivity with no internet access. Advice on a windows help forum suggested that this problem may be solved with disabling the firewall. It did not work.

On my continued quest for solutions, I found the following suggestions:

Under Properties for the Driver found in the Network and Sharing Center, click Local Area Connection, uncheck the Internet Protocol Version 6 (TCP/IPv6) option.

Click Start in the bottom panel,

type regedit in the Search programs and files box, and click on regedit in the list.

If you are prompted for an administrator password or for confirmation, type your password, or click Continue.
Locate and click the following registry subkey: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip4\Parameters\Interfaces\{GUID}. In this registry path, click the {GUID} subkey that corresponds to the enabled ethernet interface.

On the Edit menu, point to New, and then click DWORD (32-bit) Value.

In the New Value #1 box, type DhcpConnEnableBcastFlagToggle, and then press ENTER.

Right-click DhcpConnEnableBcastFlagToggle, and then click Modify.

In the Value data box, type 1, and then click OK.

I added one more step:

Locate and click the following registry subkey:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip6\Parameters\Interfaces\{GUID}

On the Edit menu, click on New, and then on DWORD (32-bit) Value.

In the New Value #1 box, type DhcpConnForceBroadcastFlag, and then press ENTER.

The Value data box remains set to: 0.

Close Registry Editor.

Note a DhcpConnForceBroadcastFlag value of 0 disables this registry entry. You can use this registry entry to prevent Windows from using the DHCP BROADCAST flag. After you set this registry entry, Windows never uses the DHCP BROADCAST flag.

Taking these steps did not render the ethernet connection with the RTL8029 driver functional either. Eventually, I disabled DHCP and entered the static IP addresses proposed on ubuntu's community documentation site manually:

qemu-emulated Window 7 I.P. address: 10.0.2.15

Gateway address: 10.0.2.2

SubMask address: 255.255.255.0

DNS Server address: 10.0.2.3

Disabling the DHCP server did not help either. At that point, I concluded that connecting to the internet was not going to work with the legacy driver. Instead, I found here that recent versions of qemu may support the rtl8139 driver, though warning was given that this option would not provide any connectivity. Despite, restarting qemu with

qemu-system-x86_64 -localtime -net nic,model=rtl8139 -net user -m1024 -cdrom './windows7.iso' -boot d windows7.img

produced a working connection to the internet, miraculously automatically. The system loaded the RTL8139C+ Fast Ethernet NIC driver from its own driver directory. The interface uses DHCP and found the correct IP addresses. I did not have to edit anything manually. The pre-installed driver can be replaced with the most recent driver from Realtek for Vista 64 and Vista bundled in the newest auto installation program (Submission ID: 1310628).

Further experimentation showed that the RTL8029 driver can be used as well. Once specified on the Qemu command line with the "-net nic,model=ne2k_pci" option, Windows 7 will load the RTL8029 driver, if it has been previously installed in the system's driver directory and configure an ethernet connection. The same will happen, when you start qemu with the default option "-net nic".

A crucial prerequisite for Windows 7 autoconfiguration to work successfully is that the RTL8029 driver must have been installed beforehand, because the driver is not included in Windows 7. By contrast, if RTL8139 is specified with the qemu command, the system loads the RTL8139C+ Fast Ethernet NIC driver pre-installed in Windows 7.

IE 8
Finally, I was ready to install the latest updates from Microsoft and test applications. I remain unimpressed with the Internet Explorer (IE) 8. This browser denies access to sites like this blog which are perfectly accessible with older versions of IE, Google Chrome, Firefox, Opera, Safari, and Seamonkey. IE 8 help does not suggest any precise actions to remedy the problem. Hence,
I opted for Google Chrome instead.

SHARED FOLDER
To access a shared folder on the host system, kqemu must be started with the following command in the terminal:

qemu-system-x86_64 -localtime -net nic -net user -m1024 -boot c windows7.img -smb /home/user/qemu_share

In order to access the qemu_share folder from Windows 7, the folder's permissions must be set for sharing on the host. In Windows 7, the folder can then be mounted following the sequence below:

Click Start in the bottom panel, choose Computer,

right-click Network and choose Map Network Drive...,

enter \\10.0.2.2\qemu_share on the folder choice line,

choose Reconnect at logon,

and click Finish.

SOUND
Qemu provides several options for sound hardware, the only option Windows 7 recognized on my setup was:

-soundhw es1370

Ensoniq's ES 1370 technology dates back at least ten years. I found an acceptable driver on the Ensoniq support site. The driver installer location can be accessed, using the following path:

E-MU Legacy Hardware support

Audio PCI

Windows 3.1 Driver Version 3.30.06.

Alternatively, you may wish to download the driver installer directly here.

The driver's setup wizard refused to fully install the software. Intriguingly, the following Windows 7 update detected and upgraded the ES 1370 driver, I needed to set the variable QEMU_AUDIO_DRV in my environment. Which drivers are supported depends on your host operating system. The options can be reviewed with qemu -audio-help. I chose alsa. Therefore, the launch of the emulator must be preceded with the comands :

export QEMU_AUDIO_DRV=alsa
export QEMU_ALSA_DAC=dmix
export QEMU_ALSA_ADC=null

Though I turned up full volume in all controls, the speaker sound remained exceedingly attenuated. The problem does not reside with the host. My Windows 2000 XP system emulated with the same qemu version produces great sound. This shortcoming still remains unresolved.

RUNNING WINDOWS 7
At this point, the complete command line to boot Windows 7 on kqemu includes the following options:

qemu-system-x86_64 -localtime -net nic -net user -m1024 -boot c windows7.img -smb /home/user/qemu_share -soundhw es1370

I recommend to save this command line in an executable shell script. The blue screen still erratically pops up during boot. The system will boot up after a couple of tries.

Addenda

I applied the same procedure to successfully install the release candidate that became available for download on May 5, 2009. Reboot does not always work during installation. You may have to go through a number of lengthy repair routines that Windows suggests and be patient. Eventually, it will work (05/10/09).
I recently upgraded Ubuntu to Karmic Koala (9.10). This upgrade comprises a new version of qemu (0.11.0), changing the available options. With the new version, I can run the emulator with two central processing units, that is -smp 2. More types of emulated ethernet cards are available. As a consequence of the update, windows 7 uninstalled the rtl8139 ethernet driver after the next boot. I had to re-install the system-provided drivers with the Device Manager (Control Panel > System and Security > System). For sound, the es1370 option is still listed following the command qemu -soundhw ?. However starting qemu with this option now fails. I am investigating (11/28/09).
The sound problem is unrelated to qemu. After I removed the environmental variables for alsa from the start-up script, windows 7 recognized the virtual ensoniq sound card and installed the es 1370 driver. The loudspeaker icon on the bottom menu bar of windows 7 turned unstruck, indicating active sound. However, no sound is to be heard yet. The environmental sound variables many need to be set differently in Karmic, possibly because Karmic appears to use pulse audio instead of alsa by default. I am investigating further (12/22/09).
If pulse audio is used for sound, the above environmental sound variables must be replaced with: QEMU_AUDIO_DRV=pa (04/15/10).
The download link on www.softwarepatch.com for the RTL8029 driver provided above does not seem to work at times. I found an alternative at www.techspot.com here (08/03/10).
Because of recently exploited vulnerabilities, the use of the RTL8029 driver is highly discouraged. You may wish to read more here: The Stuxnet Worm, Windows & The Internet (09/29/10).

Related Posts

The Stuxnet Worm, Windows & The Internet

Reduced Calorie Diet & Body Mass Index

2009-02-26T13:22:00.062-06:00

Reduced-calorie diets seem to help diminish body weight. Sacks and others (2009) report in the New England Journal of Medicine (360:859-873) that eating reduced-calorie meals with differing allotments of carbohydrates, fats, and proteins helped 645 overweight adult men and women, who stringently dieted for two years, lose on average 4 kg (roughly 9 pounds). The findings caught the attention of the national media. Tara Parker-Pope filed a report on them in her column Well for The New York Times already yesterday evening. CNN's Madison Park posted a detailed report about the study today.

The participants of the study conducted at Harvard University and Louisiana State University had body mass indexes between 25 to 40. Body mass index (BMI) was calculated dividing the weight in kilograms by the square of the height in meters (see attached converter). Diets low and high in fat and protein were personalized to provide 750 kcal less per day than the baseline expenditure determined for each person.

Statistical analysis showed that all diets tested diminished body weight noticeably, though not statistically significantly. The various types of diet did not differ significantly in the diminution of body weight. However, encouragement with counseling reinforced weight loss. More persistent encouragement, in particular the addition of physical activity regimens, could have enhanced the observed small differences in effect.

Using the converter below, you may convert your body weight [pounds] and body height [inches] into metric dimensions according to the Système International by entering the appropriate values in the blue boxes and clicking on the green boxes. If you use metric measures, just enter your data in the green boxes. Subsequently, clicking on the body mass index button calculates this value.

Addenda

Keith Devlin discusses ten reasons why BMI may constitute an unrealistic indicator of obesity in an segment broadcast on yesterday's Weekend Edition of National Public Radio (05/07/09).
According to Will Dunam's post on Reuters dated May 3, 2009, reduced-calorie diets promote longevity possibly through the expression of pha-4 related genes, the products of which are involved in the balancing of blood sugar (07/14/09).
Nicholas Wade reports in his post for The New York Times, dated Jul. 9, 2008, on results of a longitudinal research study with rhesus monkeys (Coleman and others, 2009), suggesting that reduced-calorie dieting extends average lifespan (07/16/09).
Alex Witchel's article for The New York Times posted Oct. 6, 2009, describes the preliminary results of the first scientifically controlled studies using restricted diets on participants of normal weight. The imposed dietary regimens may be difficult to adhere to, once we are left to our own whim. The studies have not advanced to that point yet. Perhaps, the most insightful result to date is the notion that keeping track of our food intake in a conscientious effort helps (10/09/09).
The discussion below gave me pause. It would be great, if this example of weight control were the result of pure self discipline (10/16/09):

On Mar. 4, 2010, Maggie Fox posted a report with the title "Your best diet? It might be in your genes" on a recent study uncovering variations in the genes ADRB2, FABP2, and PPARG that may play a crucial role in the efficacy of diet types. The products of these genes are proteins that influence the secretion of insulin from the pancreas (ADRB2), facilitate fatty acid metabolism, (FABP2), and regulate fatty acid storage and glucose metabolism (PPARG). The study was presented at the American Heart Association's Joint Conference - 50^th Cardiovascular Disease Epidemiology and Prevention - and - Nutrition, Physical Activity and Metabolism - 2010, held in San Francisco last week (Dopler Nelson and others, 2010). The results of the study are based on the examination of 140 overweight women. Four diets were randomly assigned to the participants: Atkins (low in carbohydrates), LEARN (extremely low in fat), Ornish (low in fat), and Zone (balanced). Personal genetic profiles were analyzed in roughly 100 participants, using an assay available from Interleukin Genetics Inc. Over one year, participants with the diet matching their genetic profile lost twice the body weight (5.3 precent) than the participants on mismatched diets (2.3 percent). Note that people on restricted calorie diets lose about 10 percent in the same time span (03/07/10).
Our body mass index is supposed to range between 18.5 and 24.9 to keep us healthy. Kathrine Rosman provides some interesting facts for the benefits of a little more weight in her Wall Street Journal post with the title "A Case for Those Extra 10 Pounds" two days ago. A few more pounds seem okay, as long as they are not visceral. That's my snag (04/29/10)!
According to Gautam Naik's post entitled "Gene Curtails Alzheimer's in Mice" on The Wall Street Journal today, Guarente and others (2010) provide evidence that sirtuin 1, the product of the gene SIRT1, the expression of which is induced by reduced-calorie diets, helps prevent the formation of amyloid plaques and memory loss in a mouse model for inheritable Alzheimer's disease (AD). AD mice that had additional SIRT1 inserted into their genome showed the positive results, wheras AD mice that had SIRT1 knocked out off their genome deteriorated extraordinarily fast (07/22/10).
Rick Wilking provides as with a gallery of slides published online on Reuters yesterday with the title "Big U.S." documenting obesity in America. We may wish to consult the BMI table on slide #33. As long as our BMI is in the white part of the table, we are fine (10/09/10).
Bruce Grierson informs us in his astounding essay with the title "The Incredibly Flying Nonagenarian" published online in The New York Times today about the achievements of athletes at advanced age, the physiological and genetic basis underlying such incredible performance, and the beneficial effects of physical fitness. Persistent physical exercise routines may slow aging, because the more efficient use of calories may stem the degradation of genes involved in tissue regeneration, enhancing and preserving our physical potential. Watch Olga Kotelko at age 91 (11/25/10):

According to this Reuters chart with the title "Obesity" published online Dec. 9, 2010, BMI statistics suggest that Italians and French are less overweight than U.S. Americans and Germans. Must be the food (12/16/10)!

References

Colman RJ, Anderson RM, Johnson SC, Kastman EK, Kosmatka KJ, Beasley TM, Allison DB, Cruzen C, Simmons HA, Kemnitz JW, Weindruch R (2009) Caloric restriction delays disease onset and mortality in rhesus monkeys. Science 325:201-204.
Donmez G, Wang D, Cohen DE, Guarente L (2010) SIRT1 Suppresses β-Amyloid Production by Activating the α-Secretase Gene ADAM10. Cell 142:320-332

Secure Virtual Network Computing

2008-11-17T11:39:00.054-06:00

With the ever increasing speed of internet data transfer, virtual network computing (VNC) has become a viable alternative to stand alone work stations with duplicate software installations. VNC enables users to access the desktop of a work station remotely via the internet through a slim client computer for the processing of data with heavy-weight applications. Commonly, a window server application needs to be installed on the work station and a viewer application needs to be installed on the client computer to display the work station's desktop. Once the client is connected to the server, the user can work on the server like on any other desktop. Modern VNC versions even allow drag-and-drop file transfer between client and server.

Commercial VNC software is available for Linux/Unix, Microsoft Windows, and Apple's OS X operating systems (e.g. RealVNC). Apple provides simple remote desktop connectivity for its computers in Leopard (ARD 3.2). Redstone Software offers only the Vine 3.0 server for OS X (formerly OSXvnc) free and charges for a sophisticated server/client combination with which you can drag and drop files. However, Chicken of the VNC can be used for viewing on OS X as a free complement to the Vine server.

A great advantage of VNC is that the connectivity between client and server is platform-independent. Their operating systems need not be identical. I currently use TightVNC. TightVNC is an award-winning open source project offering a highly functional package for UNIX- and Microsoft Windows-based operating systems. I installed the server on a computer running Microsoft Windows XP-64. I view the desktop on a machine running Ubuntu 8.10 or on Apple computers running OS X. TightVNC comes with a platform-independent, Java-based viewer that can be used as stand-alone application, provided the Java Runtime Environment (JRE) is installed. JRE is included with OS X and can be easily installed with the Synaptic Package Manager on Ubuntu. In addition, a viewer applet can be accessed on the server through the browser. The latter option does not require any install on the client.

VNC connections commonly use ports with numbers equal to or greater than 5900. The last digits identify the display number. The user needs to register a user account with username and password on the server. The regular login dialog is unencrypted, rendering the server vulnerable to break-ins. This hazard is perhaps tolerable on a local network protected by a firewall. However, if the server is slated to be accessible over the internet at large, unencrypted login exposes the server to excessive risk. Remote login to the server through a secure shell with RSA encryption is preferable and SSH protocols provide this option.

SSH encryption is fairly safe, particularly when passphrase encryption is employed. As additional precaution, the administrator may wish to ascertain that root login is disabled in the SSH configuration script. I once made the mistake to administer a computer with superuser privileges choosing a common English noun as password. I did not know that root was enabled by default as username on port 22 in the SSH configuration script. It took some miscreant five years to guess the password, turning my machine into a reflector for evil data transfer, until the gate keepers found out about it. To preempt complications of this nature in the future, I use sudo these days, if I need to execute commands with superuser privileges.

Remote login to a VNC server from a VNC viewer with SSH entails two steps: first the connection to the VNC server needs to be successfully established via the SSH protocol, and then the viewer has to be opened using the connection. I developed a Java-based application on the intel architecture from open sources for Apple's OS X 10.4 (Tiger) and 10.5 (Leopard) that achieves this goal in a convenient amalgamated process. This sVNC client combines jcraft's SSH login with the TightVNC viewer.

If you like to use the sVNC client, you may wish to download the zipped folder containing the application bundle with a click on the logo on the left (trademark application #77642277, pending). Copy the application to a folder of your preference. Double-click on the application's icon to launch the SSH dialog. You are asked to enter your username @ the VNC server identified by name or IP address. Subsequently, you need to provide a listening port:host:remote port combination, specifying the local host and the ports for the displays. Commonly, you can use the default combination 5900:127.0.0.1:5900. This combination specifies the host and serves from and to displays with the number 0. If the VNC server uses a different display number, e.g. display 1, change the first number to 5901. If you wish to use a different display on your home computer, you need to change the last number accordingly. There is no need to change the address for the local host.

After you provided this information, the SSH dialog generates RSA keys. Your consent is needed. You must enter your password for the VNC server. The connection is attempted. In case of success, you need to provide your password once more for the VNC viewer, the display of the remote desktop should pop up, and you are ready to go. The configuration options to the viewer can be found in the README file included with the application in the zipped folder.

The client works well for me. If the remote desktop seems unresponsive to clicks, hit the refresh button on the viewer's top panel. The viewer may crash when the server is accessed in a sleep state. Do not be discouraged. The viewer will work on the second attempt. I hope sVNC will be useful to you.

2.5-inch Hard Drive Enclosures: A Journey

2008-10-26T10:36:00.028-05:00

Hard drives have become small theses days. I needed a handy external storage device for travels with the laptop. After considering enclosure size and prize range, I opted for a Hitachi TravelStar 200GB 2.5-inch hard drive and sought the enclosure, separately. The latter turned into an adventure.

I looked up enclosures with USB-2.0 connectivity. USB cable and a protective pouch are usually included. The first I ordered from an internet retailer of national renown had a nice steel case and the option to power it with a separate supply that was not included. It cost $30.- and came with a hefty rebate offer. I never received the rebate, because the rebate center did not accept the only proves of purchase I had, that is copies of the order confirmation and the packing slip. You will not see advertisements of this retailer on my sites.

Installation of the hard drive into this enclosure was simple. You had to remove a tiny lock-down screw on each long side, pull out the front of the enclosure with the USB interface, attach the drive to the connector, slip the assembly into the case, and re-secure the front with the screws. Sadly, the drive would not start up.

The next enclosure I bought cost only $10.-. No rebates were involved. Installation was simple. Again, the front needed to be pulled out of from the case, and the drive had to be connected. This time no screws were needed. The drive was supposed to be powered by the USB connection. To my frustration, the drive attempted to start up with plenty of clicks, but did not succeed. What was the problem?

I did not read the specifications in the extended description of the enclosure. Apparently, the USB interfaces of some SATA drive enclosures are limited in the storage size of the drive they can handle. This limitation is not related to the price. I eventually found a suitable enclosure that red.populus offered on ebay for $11.-, shipping and handling included.

The top and bottom shells of Dragonext's USB-2-0 2.5-inch SATA external hard drive case must be slid backward to open the enclosure. The drive must be connected to the interface and secured with four screws to a bottom plate, and top and bottom shells are slid forward over the drive until they snap into the front. Once connected to the computer with the included cable, the drive started up without complaint.

The cable is special in that it sports two USB connectors on the computer-side; one for power and one for data. Taking up two USB connections may seem a nuisance, but constitutes a small inconvenience for a functional case that does not need a separate power supply.

Of course, if you choose to avoid any hassle and are looking for a drive that is more sturdy and extravagant, but bulkier, you may opt for the product below:

Ginkgo & Stroke

2008-10-10T14:40:00.005-05:00

Today, Oct. 10, 2008, Tara Parker-Pope informs us in a post under her New York Times health column Well on recent research suggesting that Ginkgo tree extracts may reduce brain damage after stroke. The study is published in the journal Stroke. The researchers affiliated with Johns-Hopkins University and La Fondation Ipsen temporarily blocked the middle cerebral artery in mice to induce an ischemic stroke. In mice treated with Ginkgo extract, the volume of brain tissue affected by the stroke was only about half that observed in untreated mice. The extract had no effect in mice that lacked heme oxygenase 1, an enzyme known to reduce oxidative stress caused by free radicals. These findings may encourage further basic research, ultimately opening a path for future pre-clinical and clinical trials.

I developed a liking for Ginkgo trees ever since I saw my first one on a high school visit to Heidelberg. The tree stood in the park outside the castle, solemnly holding its own one of a kind. They are not native to Germany. A lord long gone had purchased a seedling from China and planted it as a curiosity.

Ginkgo biloba is a species of ancient plants. Herbivorous dinosaurs already dieted on their broad leaves. The leaves are misleading. The trees are more closely related to conifers than to broad-leafed trees. Ginkgos are abundant in the Southeastern United States where I live. The leaves add a wonderful bright yellow to the panoply of foliage colors in the fall.

The leaves' peculiar shape caught the eye of the eminent German poet scientist J.W. von Goethe two centuries ago. Eternalized in a remarkable poem, Goethe calls to our attention the fact that the observer cannot tell from the leaves' shape whether they constitute one leaf split into two or two leaves fused into one. I quote "...Eins und doppelt bin."

This notion of ambiguity applies immediately to the discussed stroke research. As long as we do not know precisely which substances in the Ginkgo extract affect the ischemic brain tissue and what underlying molecular mechanisms reduce the impact of the stroke, medication with Ginkgo extracts is ill-advised. Particular prudence should be exercised, because the extracts are known to diminish blood clotting, exposing patients on blood thinners to additional risk.

Social Data Mining: Work in Progress

2008-10-08T11:30:00.003-05:00

Demographic data mining is a recent enterprise in social networking on the internet and appears to be still in its infancy. Social web sites commonly ask the user to create a profile to help them find friends, that is the like-minded they wish to communicate with. The collected personal information is used to define the users' context of life and attempt to match it against that of other users. In my experience, the most widely-applied data mining strategies quarry only the simplest information, e.g. residence, school or employer, and only the most fundamental personal characteristics are examined. That is, age, gender and marital status are used to draw conclusions about the interests and preferences of the profiled.

In addition to providing matches for users, these data are used to steer advertisements to the users' web pages that inform on products and services they may be interested in. This seems only fair, because the service on the web site is provided free of charge. However, the conclusions drawn from the users' profile seem at times unrefined and presumptuous, painting a startling image of the average person.

I am 53, male and married. I provide these demographics in my profiles. On my page on one popular social network site, I recently found advertisements of services entitled "Dating for Seniors" and "Meet Married Women in your Neighborhood". Welcome to the real world? A look at my profile however suggests that I am not a likely consumer of these services. Such misguided targeting may only convince users to develop avartars, i.e. virtual cyberspace personalities, bearing little resemblance with their actual lives. Credibility suffers, while professional advertising agencies must be interested in finding true clients for their customers. Obviously, there is ample space for improvement. Until more suitable targeting has been developed, we must tolerate these inconveniences with a . They are minor compared with the great opportunities social sites offers.

AquaInkscape

2008-09-30T12:29:00.035-05:00

With Aqua, Apple provides its customers with an intuitive graphical user interface (gui) and an appealing desktop display. Apple's OS X, also known as Darwin, is a UNIX-based operating system. This permits frugal users to run free open-source software on their Apple computers. The main obstacle that needed to be overcome was that much open source software is developed for operating systems that use the X11 window system, whereas Apple uses its own Quartz Compositor. Apple solved this problem with developing X11.app. This application consists of a command line terminal that is run from an X11 window server. Open-source applications that were compiled for Apple computers and use X11 can be run from this terminal. A number of such applications can be found under the Unix & Open Source tab on Apple's Mac OS X Software downloads site.

The X11 windows system does not take fullest advantage of Aqua's capabilities. The X11 project develops very rapidly, progressing faster than the three-year release cycle of Apple's OS X. Incompatibilities may result. Hence, I decided to adapt available open source software that scientists commonly use, e.g. desktop publishing and graphics applications, to run with the Quartz Compositor as back end. The applications take advantage of Aqua and do not need X11.app. They are compiled for Tiger (OS X 10.4) and, more recently, Leopard (OS X 10.5) on the intel (i386) architecture. The applications can be downloaded from my project site on sourceforge.net called "Software for Small Budget Science".

Last week, I completed my first Aqua-build of Inkscape for Tiger. I named it AquaInkscape. A version for Leopard is in preparation. Inkscape is a versatile vector graphics application with a great number of tools and options similar to Adobe's Photoshop®
and Illustrator® CS4
or the open source application GIMP. These applications feature freely-floating tool bar, control and navigation windows that are detached from the main window with the canvas. This is great when you have two monitors attached to your computer and are able to reserve one for the tools, control and navigation, leaving the view of your work unobstructed on the other. However, if work needs to be done on a small display like a PDA, clutter ensues and free-floating windows quickly finish unretrievable in the stack of open windows.

By contrast, in Inkscape the tool bar is integrated into the main window. This may seem old-fashioned. But I found it truly helpful when working in a confined space. Detached popup windows provide further options. But these windows are task-specific and do not need to be left open. This design profoundly enhances the user's experience of an uncluttered workspace.

To software engineers, Inkscape offers the advantage of an application that can be accessed by and integrated into programs written in Python, a programming language widely used in science and for internet applications (e.g. the Google App Engine).

Below, I briefly summarize my experience with a number of issues pertinent to building AquaInkscape.

I compiled 35 packages to build this application. The needed packages were identified using configure --help, beginning with Inkscape and proceeding through the packages required subsequently. I used the latest stable releases for most packages and developmental versions for gtk+-2.0 and cairo.

For compilation, I followed GNU standard procedures, that is the commands aclocal, autoconfig, autoheader, and automake, if the package was unconfigured. Sometimes, packages provide an autogen script that can be used instead. Then, I execute configure, make, make test and, if all goes well, make install. I install into the default location, that is /usr/local.

In order to be able to run the packages with Aqua, the compiles have to be configured without X11 and with quartz, if configure --help indicates these options. The packages gtk+-2.0 and cairo are instrumental for rendering Inkscape's gui. They can be built for quartz.

The demos in gtk+-2.0 and tests in cairo instruct about the potential performance of these packages with Aqua. In the demos for gtk-2.0, not all options worked because the default collection of theme icons known as hicolor icon theme was incomplete. I found a full set that fulfilled the necessary requirements in the Mac4Lin Leopard gtk icon theme package.

With my cairo build (version 1.7.4) a third of the tests failed. Most failures involved tests with very large font and some svg options. Only few failures were associated with quartz. I decided that the chances were good for a functional version of Inkscape for Aqua.

It is important to configure the font usage accurately, because faulty instructions may result in ugly font output and Inkscape crashes. I use fontconfig and freetype2 for font management. On my system, the path to the default fonts needed in the configuration of fontconfig is /usr/local/etc/fonts. You may furnish additional font with ghostscript or gutenprint. Both packages will install fonts in /usr/local/share by default.

For ImageMagick, I used precompiled binaries.

After all required packages had eventually been installed, I was prepared to compile the inkscape binary with the MacOSX build script provided in Inkscape's packaging folder. I specified the options openoffice files, internal perl, LittleCms, Inkboard, Poppler-Cairo, ImageMagick/Magick++, Libwpg, as well as with and without internal python. For Python, I compiled and installed version 2.5 on my system. Precompiled MacOS X binaries of Python 2.5 cannot be used because they are incompatible with the Numpy package that Inkscape requires. Therefore, the python version of AquaInkscape contains Python 2.5 modules.

Finally, I had to edit the scripts for packaging so that AquaInkscape would launch without asking for X11. The resulting application seems in good working order. I have not tested all possibilities. The only shortcoming I found so far and have been unable to resolve is that the program does not respond to the quit button in the application menu once inkscape-bin is running. The application can be quit, however, with the button in the pulldown menu under File and with the button in the icon menu on the dock. Clicks with the mouse expand the compressed tool icons in the side bar. Have a try and be patient at start up! The application takes some time to launch.

Subprime Lending: Truthiness & Delusion of Mind

2008-09-26T12:59:00.008-05:00

After Citigroup collapsed late last year, I wondered whether the underestimation of risk in securitized investments was the result of a failure in using scientific methods of risk assessment. Read my post dated Dec. 5, 2007. Soon, I learned that such methods were widely employed in investment banking and insurance. Apparently, misconception prevailed about the reliability of the predictions that these methods provide.

A year on, we find out that the statistical software did not fail. Any scientific method will provide a realistic prognosis only, if the entered data are correct. We presume that a loan is approved with the assurance that the applicant is going to be able to service the payments through the loan period, provided that the applicant's financial situation does not change dramatically. This good practice seemed to have been widely ignored in the subprime and prime ARM (adjustable rate mortgage) loan market of the past eight years. Too often, applicant naivety and lender carelessness took sway over prudence.

In hindsight, the considerate risk analyst would have welcomed an additional column in the input spreadsheet with a truthiness factor weighting each loan. That is, the loan agent is asked to assign a score estimating the applicant's ability to service the loan once the interest rate resets. The agent could be rewarded for good judgment and the improvement in the quality of loans issued. Perhaps such estimates from the field would have helped to predict the risks with the ensuing securities more accurately.

In the absence of such information, the analyst is left to look at the foreclosures on the loans to examine risk retrospectively in the hope that forward projections gain accuracy. Paul Jackson reports on housingwire.com in a post dated Sep. 5, 2008, that the Mortgage Bankers Association (MBA) announced in early September that more than 9 percent of U.S. mortgages were delinquent or in foreclosure at the end of the second quarter of this year. “Subprime ARM loans accounted for 36 percent of all foreclosures started and prime ARMs, which include option ARMs, represented 23 percent,” Jackson quotes MBA's chief economist Brinkmann.

Addendum

As The New York Times reports today, I am not far off with my assessment. Listen to this incredible story (11/5/08).

Cell Phones & Brain Cancer

2008-08-21T12:51:00.024-05:00

Remember Jon Krakauer's account of tragedy on Mt Everest? We have an irrational perception of acceptable risk. For sheer enjoyment, we frequently are willing to expose ourselves knowingly to the possibility of severe injury. Just three weeks ago about a dozen climbers succombed to an ice fall on K2. In any good summer, roughly 100 people perish in mountaineering accidents in the Swiss Alps alone. V. Lischke and others report 462 fatalities for 1997 Europe-wide. According to the U.S. Dept. of Transportation, 4,810 Americans perished on motor bikes in 2005, translating into 73 riders per 100,000 bike registrations. The fatalities continue to rise. Last year 5,154 Americans died on their bike.

By contrast, according to wrongdiagnosis.com every year fewer than 50 in 100,000 Americans battle glioblastoma multiforme, that is the most aggressive and fatal type of brain cancer. Recent research studies link brain cancer with electromagnetic radiation from cell phones. A discussion of some studies can be found here. The debate has become ever more intense since Senator Edward Kennedy was diagnosed with malignant glioma, probably a glioblastoma multiforme. I discussed this type of brain cancer in my post dated Jun. 3, 2008.

Prudence is advised in the debate on a possible association between cell phone-related electromagnetic radiation and brain cancer. The energy of electromagnetic waves is directly proportional to their frequency. The spectrum of the frequencies ranges from meter-long radio waves via the nanometer-long waves of visible light to atom-sized gamma rays. The shorter the wave length, the higher the frequency and the greater the energy that potentially can harm our genes. DNA may be altered directly through the absorption of radiation energy or indirectly through radiation-produced free radicals that may react chemically with the DNA. Even low amounts of energy can damage DNA and may theoretically result in uncontrolled cancerous cell proliferation. However, our cells are provided with DNA repair mechanisms. Commonly many hits are needed to overcome the defenses and cause noticeable damage. Alas, the defenses appear to wear out with advancing age and we may become more susceptible to cancer.

To determine thresholds for harm from electromagnetic radiation, epidemiological studies are conducted that compare the health records of people with known exposure to those of people who match this group in all aspects, except the exposure. Professionals with job-related extraordinarily high exposure are frequently enrolled in the group of the potentially affected to improve the chances of discovery. Hardell and others (2005) observed that Swedish long-term users of analogue mobile (NMT) phones like the now obsolete car phones developed a significantly higher risk for auditory nerve tumors (acoustic neuromas) on the side of phone use. Digital mobile (GSM) phones like our modern cell and cordless phones did not increase this risk consistent with findings in Denmark (Christensen and others, 2004). The prevalence of acoustic neuroma in the U.S. is low. Less than 1 in 1000 Americans is affected. According to ctia.org, roughly 264 million wireless subscriptions are active at present.

Though the results of such studies provide important leads on the type of the potential harm to our health and may be instrumental for safety considerations at the work place, they cannot be easily extrapolated to regular phone users. The amount of radiation energy deposited in the auditory nerve can only be estimated. Moreover, the effects cannot be ascertained with the data collected at higher doses because of the non-linearity of the relationship between dose and effect and the increasing scatter of the observations at progressively lower doses. The concerned may opt to use plug-in extensions for ear and mouth pieces or voice-activation as precaution.

Addenda

NPR's All Things Considered ran an interesting update on this issue today. You may wish to read the report and listen to Allison Aubrey's podcast entitled "Doctors Urge Research On Cell Phone-Cancer Issue" (09/25/08).
On Aug. 24, an ice fall swept away a party of twelve on the Mont Blanc du Tacul, a smaller brother of the White Lady. Eight are missing. I once stood on this mountain at a different time in the year. I was fortunate to have a friend as guide who understood risk well (10/21/2008).
Maggie Fox reports in her post with the title "U.S. senator promises look into cellphone-cancer link" published online on Reuters today that the chairman of the Senate Committee for Health, Education, Labor and Pensions Senator Tom Harkin plans to encourage more research to examine whether the use of cellphones may cause cancer. The fear persists (09/14/09).
The environmental working group lists the head-absorbed power [W/kg] of the radio waves emitted from a number of cell phones in this table (06/16/2010).
Volkow and others (2011) report in this week's issue of the Journal of the American Medical Association that cell phone radiation is statistically significantly associated with an acute increase of glucose metabolism in the temporal and frontal lobes of the cerebral cortex by, on average, 7.2 percent on the side the active, but muted, phone was held for 50 minutes. The researchers used the [¹⁸F]fluorodeoxyglucose method and positron emission tomography (PET) to determine regional cerebral glucose utilization rates in 47 healthy volunteers. They conclude in the abstract of their communication that “this finding is of unknown clinical significance.” Indeed, the finding does not provide any insight into the cellular mechanisms underlying the observed increase. Under healthy physiological conditions, brain glucose metabolism does not rise to levels posing a health hazard. I have written on this topic in my post with the title "Good News for Brain Energy Use" published Sep. 12, 2009. If the sole aim of this study had been to investigate the potential influence of modern-day cell phone radiation on brain energy metabolism, the study could have been conducted with mice at lower cost, sparing the participants unnecessary exposure to ionizing radiation from PET (02/24/11).
Epidemiologists have recently come to discrepant conclusions on the risk of cancer associated with cell phone use. According to Scott Hensley's report with the title "Cellphone Use May Be A Cancer Risk After All" on National Public Radio's All Things Considered today, a recent World Health Organization review conducted by 31 experts from 14 countries found sufficient evidence that may support a correlation between cellphone use and gliomas and acoustic neuromas. The study will be published in the July issue of the journal The Lancet Oncology. By contrast, a comprehensive case–control study with 2708 glioma and 2409 meningioma cases and matched controls from 13 countries published last year by The INTERPHONE Study Group (2010) could not establish any elevated risks with mobile phone use with certainty (05/31/10).

References

Christensen HC, Schüz J, Kosteljanetz M, Poulsen HS, Thomsen J, Johansen C (2004) Cellular telephone use and risk of acoustic neuroma. Am J Epidemiol 159: 277-283.
Hardell L, Carlberg M, Hansson Mild K (2005) Case-control study on cellular and cordless telephones and the risk for acoustic neuroma or meningioma in patients diagnosed 2000–2003. Neuroepidemiology 25: 120-128.
The INTERPHONE Study Group (2010) Brain tumour risk in relation to mobile telephone use: results of the INTERPHONE international case–control study. Int J Epidemiol 39: 675-694.
Volkow ND, Tomasi D, Gene-Jack Wang G-J, Vaska P, Fowler JS, Telang F, Alexoff D, Logan J, Wong C (2011) Effects of Cell Phone Radiofrequency Signal Exposure on Brain Glucose Metabolism. JAMA 305: 808-813.

The Body Surface Area

2008-07-16T12:47:00.073-05:00

Today I deployed my first Google application developed with the Google app engine (GAE). You have to be connected to the internet and establish an account with Google to use it. BSA Calculator is a body surface area (BSA) calculator with a simple layout in font that the visually challenged like me can read without squinting.

The large font may also be helpful to the busy physician using an iPhone.

Physicians regularly use body surface area to calculate drug doses for their patients. Notably, the dosing of chemotherapeutic drugs like Cisplatin is determined according to BSA. I use it to control my diet. When my BSA climbs over 2.1 m², it is high time for me to act.

Estimating human body surface is not trivial. It is difficult to develop one formula that fits all, simply because we come in a wide range of shapes and sizes and the relationship between body height and body weight changes considerably while we grow up. Medcalc provides a handful of methods and references. The suggested formulas have been empirically derived from measurements of height and weight. J.D. Current, 1998, and Than Vu, 1999, critically review the methods in use. Boyd, 1935, published one method that is still widely employed. In recent years the method of Mosteller, 1987, has gained popularity. Both are available in BSAcalc.

References:

Boyd E (1935) The growth of the surface area of the human body. University of Minnesota Press, Minneapolis.
Mosteller RD (1987) Simplified Calculation of Body Surface Area. New England Journal of Medicine 317: 1098.

The BSA calculator accepts only metric dimensions (Système International). You may use the converter below to convert body height and weight from inches and pounds to centimeters and kilograms, respectively. Enter the body height in inches in the inches box and click on the centimeter box. In analogy, enter the body weight in lbs into the pound box and click on the kilo box:

Proceed to the BSA calculator.

Addendum

A authoritative study by Sacks and others (2009) in the New England Journal of Medicine (360:859-873) demonstrates that we can diminish our weight on reduced-calorie diets (02/26/09).

Who Owns Your Site?

2008-07-10T12:44:00.005-05:00

Recently, a friend of mine was pondering some existential questions about websites. She had started an own business selling customized arts and crafts over the internet. She found an appealing name for her site. She tells us on the site, how her real business grew from an unreal idea. The site offers to customers options to create the product they like to purchase. Sales have been taking off. The business shows all signs of success.

My friend is not a web expert and outsourced site development and administration. Now, questions arise about the ownership of the site. The details of site registration have fundamental implications for ownership, particularly when the site name is synonymous with the brand and the businesses vitally depends on brand recognition. The person who registers the domain name of the site with the registrar may actually own that name. This person may also own the administrative privileges to add and delete data and permit other parties access to site administration. The administrator can potentially read the e-mails on the site.

For those who are not savvy in the use of the terminal command line, I found an easy way to obtain information about your site. Go to Broadband Report's whois browser application, type in your domain name, that is the site name without "www.", and hit return. Check carefully that the entries match your expectations.

How to Build a Silent PC: Episode II

2008-07-02T12:41:00.054-05:00

In episode I, I narrated my first experience with building a personal desktop computer, describing important discoveries I made in the process. Earlier this year I build my second PC. Below I shall summarize the essential lessons I learned.

The purpose of this project was to provide my son with a powerful computer to use for his graphics and arts work as well as his games. My son is learning animation with Maya Learning Edition and chess with Chessmaster.This computer would be stationed in his room. The first machine I built boasts 7 fans. The noise these fans collectively produce is considerable. You would not want to have them in your bedroom. Therefore, I was looking for a very quiet design.

On my search for the noiseless machine, I happened on cases for home theater personal computers (HTPC). Computers for the control of high fidelity entertainment centers must be designed noiseless. Compared with standard cases, the fans in these cases commonly run very quietly. As a drawback, I did not find a case with a small form factor. Even the cases, that can house micro-sized (mATX) motherboards, are not much smaller than mid-size towers. At the high end of the price range I are the impressive cases by Zalman. I opted for Antec's Fusion V2, which is comparatively low in price.

Owing to my good experience with Micro-Star International and Advanced Micro Devices, I chose the K9MM-V motherboard and an Athlon 64 3500+ Socket AM2 central processing unit (SKU:CP2-AM2-3500 AV). I purchased both in a bundle for a bargain price from TigerDirect. The "V"s are important. They indicate that board and processor are designed to fully support virtualization. Microsoft's Hyper-V virtualization software is under development and can be tested with a compatibility check utility that AMD provides for download. I opted for Microsoft's Windows XP Professional x64 as operating system to let my son run his applications smoothly and perhaps will use Xen for virtualization.

In my first assembly, I had the CPU pre-installed. This time, I mounted it myself, before placing the board in the case. It was straight forward. A close check of the instructions in the MSI manual sufficed. You must ground yourself, before you start! The designated corner of the CPU and has to match up with that of socket on the board. When the chip is properly aligned, the pins slip smoothly into the holes of the socket. No force is needed. The chip is fastened by pushing down a lever. Then, I applied an even layer of thermal grease and mounted Ultraproduct'son top. The fan has a diameter of 92 mm and provides air flow at 53.5 CFM. The motherboard was ready to be placed in the case.

Next, I filled the the board's memory slots with a pair of PNY 1GB memory modules MD2048KD2-667.
Before you order the memory modules, validate the specifications with the recommendations on the manufacturer's site.

The case came with a 430-W power supply with all necessary connectors attached and provided two internal and one external drive bay. I loaded one internal bay with a Seagate SATA hard drive. I used a Barracuda 7200.10 320GB drive.

The SATA cable came with the motherboard. The board features a second SATA input. I may install another SATA drive in the empty bay. The external drive bay was reserved for the DVD/CD drive. I installed Mad Dog's multimedia IDE drive and connected it to the motherboard with the round cable included in the MSI kit.

Micro-ATX motherboards are about 2" shorter than the regular ATX boards, but still accommodate standard PCI and AGP cards. This was important to me, because I had PCI cards in stock and wanted to be able to use AGP graphics cards. I swapped out the ATI Radeon 9600 PRO card used in my first project for a 9800 series card and installed the old card in this computer. The reason was an important discovery. If you use a mATX motherboard, ensure that the graphics card you intend to use matches the shorter length of this board. The ATI Radeon 9600 card was sufficiently short.

The motherboard accommodated all connections from the case, except the firewire port. Therefore, one of the three PCI slots had to be reserved for a USB/firewire card. I purchased Ultra 8 Port USB 2.0/Firewire PCI Combo Card.
Another slot was taken by a wireless network adapter card. I chose Cisco's Linksys WMP54G PCI Wireless Adapter v41.The Linksys software does not function properly with Windows XP x64. A compatible driver for the card can found at Ralink support. The driver that matched my card can be downloaded at PCI/mPCI/CB(RT256x/RT266x). After insertion of the PCI card, restart the computer and the hardware wizard will announce that new hardware was detected and prompt you for a driver. At this juncture, the wizard has to be manually pointed to the RT61 file in the Ralink folder under program files (x64). If the wizard does not prompt you, the driver can be loaded manually for the network adapter in the device manager. Should Ralink terminate its support for XP, the driver I used is available for download at BMI under downloads.

As useful amenities, I added Gyration's gyrotools.The compact keyboard is light-weight and short. The buttons have a good feel. The optical air mouse doubles as remote control.

You may wish to check out this book, if you seek more advice. I own a few books in this series and found them helpful.

Build Your Own PC: Episode I

2008-05-22T11:26:00.051-05:00

I grew up in an era in which kids learned basic car maintenance so that they knew how to fix problems on the road. I have worked with computers daily for decades, and I found it more and more vexing that I did not have the slightest idea of what was going on inside them. I could not even point out the vital parts of the creature, once the abdomen was uncovered. One day four years ago, I decided that the best way to find out was to construct one from parts. Below, I summarize this experience.

The process was actually fun. The computer cost about $1,500.- including the extras. These days, a turn-key PC sells for as low as $200.-. Of course, that version does not come close to the performance of the computer I built and a bare-bones PC with a comparable performance would cost at least $500.-. I can only speculate whether that system would be equipped with the same quality parts. The consideration is similar to buying a home in a new development. The base price for the home may be economic, but the quality of the piping, the fixtures, the flooring, you name it, will be poor. You may look at early and costly repairs. Of course, you are offered pricey upgrades. I opted for quality parts. To keep the running costs low, I originally chose a 350 W power supply, but upgraded to 430 W after I switched to a more powerful graphics card and increased the memory. Assuming that the computer is consuming half that wattage on average, the monthly expense for electricity is about $11.- in our utility district. We have got three computers in our household that essentially run 24/7. The expenses add up accordingly.

It may seem dated to describe my first experience with computer assembly four years after the fact. Earlier this year, I build my second computer incorporating the lessons I learned from the first. Hence, I sum up the earlier experience in a first installment and will post important aspects of the most recent experience in a second installment.

I wanted a machine that could process great amounts of imaging data fast. I chose an MSI K8T MASTER2-FAR motherboard that accommodated two 64-bit AMD Opteron 242 central processing units (CPUs). I opted for this architecture at the time because it is compatible with Sun Microsystem's Solaris 10 operating system and I needed to run an application that was specifically compiled for Solaris. The latter is doubtlessly an advanced operating system geared toward servers supporting large networks and by far exceeds my small purpose. I found it more appropriate eventually to run the machine with Ubuntu's Linux-based operating system (currently version 8.04). The latter provides up-to-date Gnome and KDE desktops and the recent upgrades have proved flawless in execution and very stable.

Unfortunately, my CPUs are not yet the type that is enabled for virtualization. Thus, running virtual machines is not as efficient as with more recent AMD CPUs (look for "-V" in the name). Despite this shortcoming, I am running Microsoft's Windows XP on QEMU with satisfactory results.

I bought most parts from internet vendors. At the time, I did not feel confident enough to mount the CPUs myself and had that done for a small fee by Spartan Technologies from whom I purchased the motherboard, the processors and coolers. Everything else I put together. The assembly did not require a great level of experience and was straight forward. Following the instructions provided in MSI's motherboard manual was a safe route to success. An important rule to obey is ensuring that you are not statically charged by electrically grounding yourself before you touch any parts. That is, tape the exposed end of a copper cable to the skin of your arm with medical tape and wrap the other exposed end tightly around a water faucet in your kitchen or bathroom. Waterlines provide the best connection with the ground. At first, I was confused by the great variety of power supply connectors and the many different adapters. Closely studying the drawings and photographs in the MSI manual straightened out the misunderstandings. My choice of hard drives, the memory, the CD/DVD writer and the graphics card was guided by the list on the motherboard retailer's website of items that other customers bought with the motherboard. In the meantime, I have developed a few own preferences:

I chose a Cooler Master Mid-T tower caseand a Thermaltake PurePower 430-Watt power supply.

I prefer Kingston Technology memory for RAM. I purchased HyperX DDR400 PC3200R memory (KRX3200AK2/1G) for this machine when I built it, but expanded with Kingston HyperX 2048MB PC3200when the price fell. Before you order the memory modules, validate the specifications with the recommendations on the manufacturer's site. The safest way for selecting RAM modules that work with your CPU/motherboard combination is to visit the support sites of your CPU and motherboard manufacturers and choose from their lists of verified products. In addition, I learned the basics about DDR memory here and found the MSI user's forum informative.
After three years, I decided to increase RAM since prices have fallen considerably. MSI provides conflicting information about the maximum RAM size. After studying the MSI forum ad nauseam, I came to the conclusion that between my Opteron 242 CPUs and the motherboard, the largest memory I could install was 4 x 1-Gb. I eventually settled on Corsair 1GB 184 DIMM PC3200 ECC DDR RAM (CM72SD1024RLP-3200). The motherboard requires that the modules are registered and unbuffered. Corsair support was very helpful in choosing the correct kind. The installation, seemingly straight forward, turned into an ardourous journey. My first idea was to add two Corsair modules to the two 512 MB Kingston modules I already had. Mixing different brands, types and sizes did not go down well at all. The machine would not boot and stuck with emanating regular beeps of distress. Removing the Kingston modules solved this problem. However, when I filled the now vacant slots with two more 1-Gb Corsair modules, the bios post hung at the memory check. My bios version is 1.3b5. I first thought I needed an upgrade. After reading a number of harrowing stories about the possible consequences of flashing your bios (one guy ended up taking a kitchen knife to the motherboard), I decided to keep this option as the very last resort. I spent a whole week on shopping for solutions, trying to change all sorts of bios settings. In the end, I actually did not need an upgrade. The solution was (i) to re-install the old Kingston modules, (ii) set the AGP aperture in the bios as low as possible for the system to work (64M), (iii) shut the computer down, (iv) swap the Kingston modules with the four new 1-Gb modules, and (v) restart the computer. Voila, the memory check finally worked. With further experimentation, I found that the AGP aperture setting dictates the size of RAM the bios is willing to check. 256M AGP aperture permits 2Gb; 128M 4Gb. So my final AGP aperture setting is 128M. The RAM size count stops about 500 Mb short of 4 Gb. Again searching the MSI forum, I learned that the mismatch between physically available memory and bios count is buried in the design of the bios. I decided to content with what I gained.

I prefer hard drives manufactured by Seagate Technology. They really served me well over the years. One is 15 years old and still works without fail. I decided to use two Barracuda 160GB SATA drivesmirrored in a RAID for user data. Both drives could be connected directly to the motherboard. A 400GB ATA drivestores the operating system.

I chose Pioneer's DVR-116 DVD/CD drive,perhaps because my first hifi headset was made by Pioneer.

Unfortunately, the motherboard does not provide on-board firewire connectors. If firewire connectivity is desired, a PCI card needs to be installed, taking up one slot. Because I needed an internal recepticle on the card, I used StarTech's PCI1394_4 4-Port FireWire PCI Card.

I prefer the more expensive flexible rounded cables to connect the ATA hard drive and the DVD/CD writer to the motherboard over the conventional flat cables. An example is shown below:
The cables are sold in lengths that match the distances in the case and are more flexible and thus easier to bend. A rounded ATA cable came with the motherboard. However, it was too long and stiff. I bought an extra set from directron.com.

The motherboard provides two SATA connectors for storage devices. I used them to configure two harddrives in a raid. I found the SATA connectors to be flimsy. While moving the computer to another room, the cables shook lose on the motherboard side, resulting in BIOS errors at boot time. I felt tempted to glue them in place with Silicon glue, but refrained.

The MSI motherboard has four slots for PCI cards and one for an AGP graphics card. I chose a ATI Radeon 9600 PRO card.This card provides excellent support for a 19" LCD at 1024 x 768. I recommend to use a more upscale card for larger screens and higher resolutions. Moreover, I strongly recommend to ensure that the card of choice fits in the case before it is ordered. The Radeon 9600 card is comparably short. Upscale graphic cards can be quite long and the available space in the case must be shared with the thick cables connecting the hard drives and other peripherals to the motherboard. Room may run out even in a standard case for ATX-size motherboards. I eventually upgraded to a ATI Radeon 9800 SE.

The heat sink on the Radeon 9600 graphics card processor and the fan are small. I decided to replace them with a copper heat sink and a larger fan I found at CompUSA (VGA Cooling Kit SKU 336044). The heat sink slightly egged on a capacitor on the card. I ground an indentation into the sink large enough for the components not to touch. Though rough, the solution improved cooling remarkably. Before mounting the heat sink on the processor, it is important to apply a thin layer of the thermal grease provided with the kit evenly on the chips surface. The kit came with a number of additional small heat sinks that can be glued on various integrated circuits on the board. I mounted them all, thinking that every little bit helps. The kit cost less than $20.- and visibly stabilized the cards performance. The screen does not flicker anymore when its hot. Finally, leaving the PCI slot next to the graphics card open considerably boosts heat dissipation.

Running a wireless internet adapter PCI card on Linux is problematic. Before you order, ascertain that the manufacturer supports your operating system or, at least, that the drivers are available for the card of your choice. I have had satisfactory experience with D-Link's DWL-G520.Instructions for the installation and the configuration of the driver can be found here. By contrast, NETGEAR does not provide any support for Linux. An open source project develops Linux drivers for ACX-111 chipset-based cards (e.g. NETGEAR's WG311 PCI Wireless Network Adapter.However, compiling and inserting kernel modules is needed, potentially interfering with your next kernel upgrade. You may be facing a new install instead of an upgrade.

I filled the last remaining PCI slot with the Creative Labs Sound Blaster Audigy SE sound cardand equipped it with JBL's Creature II.

Finally, I can give some advice on fans. My computer houses seven fans, two in the case's front, one in the rear, one in the power supply, one on each CPU and one on the graphics card. As a result the machine vacuums up air-borne dust with remarkable efficiency. Once we had construction in the house several rooms away. An old wall was broken down. I forgot to turn off the computer on that day. The fine, invisible dust from the construction ruined the bearings of the spinners on the CPUs and the graphics card as well as the large fan in the case's rear. The fans did not stop spinning, but could not maintain the required speeds to cool the CPUs effectively. Temperatures reach more than 100° F (38° C) outside for several weeks over the summer where we live. It does not cool down much during the night and attempting to save on the electric bill exposed the CPUs to the risk of overheating. In this situation, the Cooler Master's Aerogate II fan and temperature control unit I had installed in one of the external bays in the case's front warned me of the impending disaster and saved the CPUs.

I installed magnetic levitation fans and one of Y.S. Tech's tip magnetic driving (TMD) fans as replacements. I used teflon screws to mount the TMD fan. It is fabulous. A good demonstration of these fans can be found here.

You may wish to check out this book, if you seek more advice. I own a few books in this series and found them helpful.

The Value of Education

2008-01-24T11:09:00.012-06:00

My son attends Meigs Magnet School since last fall. The school is part of Metropolitan Nashville Public Schools. Acceptance is based on grades and a lottery. The other day, I had the chance to visit during class. I received my primary and secondary school education in Germany. The visit was my first at a US magnet school. The teachers were competent and involved. I saw children of a great variety of ethnic and racial backgrounds. Their diverse faces shared one striking feature: big, bright, and curious eyes. I was deeply impressed by their ubiquitous thirst for knowledge and cannot imagine a more profound demonstration of the essence of education. Its success cannot be bought. Rather, a successful education depends upon opportunities given and opportunities taken.

This principle also applies to higher education. Currently, graduating high school students and their parents are preparing college applications in this country. The New York Times ran an instructive article about the work of guidance councilors on January 4, 2008, accompanied by an illuminating Q&A section. Obviously the parties involved are deeply concerned with the wisdom of their decisions. Often, the equivalent value of a home may be spent.

I have studied and worked at four academic institutions, that is the Johann Wolfgang Goethe University Frankfurt a.M., Germany, the University of Tennessee, Knoxville, TN, USA, the University of Lausanne, Lausanne, Switzerland, and Vanderbilt University, Nashville, TN. The most demanding undergraduate class I took was a three-quarter course in Biochemistry for premedical students at UT. Regardless of private or public institution, only a select few students I met were eventually accepted by medical schools affiliated with the Ivy League. Besides impeccable grades, the most influential factor seemed the MCAT score. Everybody felt the need to take the Princeton Review.

Although higher education costs a lot in the US, it is a commodity without warranty. Education is not like a garment made to measure. Keeping with this analogy, the garment's fit depends on the ingenuity of the tailor and the wearer alike. The prospective student must decide in which environment her/his development may benefit the most. This may not necessarily be at the most expensive and prestigious school. A superb teaching environment does not depend on prime recreational facilities and posh accommodations. There may be value in simplicity. One of the most brilliant teachers whose thoughts are still remembered after more than 2,400 years worked with very little overhead.

Addendum

Unigo is the most informative college survey site, if you do not know which direction to turn (added 09/21/08).

About People with Visual Disability & the Usefulness of Braille

2007-12-09T10:35:00.009-06:00

I had the privilege to meet people with severe visual disability in a study I conducted to examine the brain regions involved in Braille reading by touch. I am interested in brain plasticity. That is, I am studying the brains ability to reorganize after a change in input from the sensory periphery, e.g. the loss of eye sight. With positron emission tomography and functional magnetic resonance imaging, brain regions can be visualized that are activated during the exposure to sensory stimuli or the execution of tasks. The findings of our study are described in detail in “Blindness and Brain Plasticity in Navigation and Object Perception”.

The study was a great learning experience for me. I had not met people with disabilities personally before.

I was very impressed with the ease with which most coped with the challenges of daily life, e.g. crossing a street. Next time when you stand at the lights of a busy intersection, close your eyes and decide when to go. You will understand immediately what I am saying.

I met many professionals with college degrees raising families. I asked one youngster whose parents are both blind, how they could possibly manage to keep track of him and his siblings. I have trouble with mine and I can see. He answered flatly: “It has never been a problem. My parents always know where we are and what we are up to.”

People with disabilities possess great wisdom of life. I once walked down the sidewalk with a participant, returning from a study. We were joking around fueled by the elation that we had a good session. The scanner had worked and nothing else had failed that day. A young man came up. He looked sullen. He called out: “Excuse me! May I ask a question? “ “Yes,” my friend replied. “I see you walking down the street a blind man. Yet, you seem happy and content. I can see. But I am depressed. I have got to take pills to get me through the day. There is no happiness in my life. Would you trade with me?” My friend answered: “Of course, I would rather see. But I learned to live without. Take a day at a time and make the best of it.”

I was impressed with the physical abilities of people with visual disability. Many actively pursue sports. There are skiing programs in Colorado like Foresight that take people with visual disability to the slopes on weekends. Small groups are assigned to sighted guides whom they follow down the trails. I am a passionate skier. When I heard of the program, I imagined something quaint. Not at all! I happened to meet a group once on a black diamond tackling the moguls at awesome clip. I still cannot fathom, how you acquire this level of skill. My performance drops considerably in the fog. I was so impressed!

They have got great talent for music and dance. Many are commercially successful. Just to name a few: Andrea Bocelli, Ray Charles, Stevie Wonder, Jose Feliciano, The Blind Boys of Alabama, Ginny Owens, and La Singla.

Finally, I learned about the meaning of Braille. A stencil with Braille cells is as essential as a notebook. The software that translates text to voice is improving at rapid rate. All major operating systems offer assistive technologies these days. Doubt is cast on the necessity of Braille in the future. However, many professionals prefer to use electronic Braille displays to be able to read the output from their computer most expediently and efficiently. Unfortunately, the gadgets are expensive compared with the voice programs, and not every employer is willing to make the investment.

Another problem is printing or, more correctly put, embossing. When I carried out my study, we attempted to emboss consent forms in Braille. It was a very cumbersome process. The software we had was rudimentary at best. The printer was highly mechanical and slow. The Braille dots were embossed in thick paper with metal pins. The noise reminded me of a machine gun and was quite unnerving. The gadget was huge, unwieldy and too expensive for personal use. I do not blame the engineers. They came up with a solution that is solid and works. However, in our day and age there must be faster, quieter and more affordable ways to emboss Braille.

I can only encourage development. Apart from the fascinating question how our brain processes Braille, Braille is here to stay for practical reasons. Who wants to depend on a computer during a black out?

Addendum

Driving a Prius people with visual disability may have great difficulty in hearing you coming. Please be considerate (09/27/08)!
New touch screen technology promises useful assistive technologies for smart phones. In his article published on The New York Times site on Jan. 4, 2009, Miguel Helft describes the work of T.V. Raman on applications tailored for people with severe visual disability using google's Android smart phone operating system.

The World Wide Web & Access for Everybody

2007-12-07T10:08:00.014-06:00

The steam engine and iron rails transformed our lives profoundly 200 years ago. The USA would be a very different place today without the railroad. I believe that the personal computer and the internet will transform our lives in even a more comprehensive fashion than the railroad. Unprecedented in our history, these tools enable us to exchange information globally and interact with each other almost instantly. In open source computing, people work together who live in Minsk, Mumbai, Shanghai and San Francisco, as if they sat across the street from each other. The resulting synergism is unprecedented in our history. On this post's day, SOURCEFORGE.NET hosts 164,138 projects and has 1,744,123 registered users.

It is essential for our future that everyone has access to the internet. Particularly, people in developing countries must not be left behind, if we wish to lessen the gap between the rich and the poor nations. That is why I encourage everybody to participate in the One Laptop Per Child (OLPC) initiative.

Nicholas Negroponte, Professor at M.I.T., launched this initiative in 2002 with the idea to create a laptop computer affordable to people with little means. This computer was conceived to cost about $100.- and work anywhere, providing access to knowledge about our world to youngsters even in the remotest areas. The product ended up costing twice as much as initially hoped.

However, there is a way to ease the financial burden. When you participate in this initiative until Dec. 31, you purchase for a bit more than $400.- one computer for yourself and another for a kid somewhere out there who has hardly ever seen a sleek gadget like this before. I bought one for my ten year-old son. Imagine two sets of eyes beaming with curiosity and excitement, one set here and a second set somewhere else, when the kids unpack this wonder machine! Join in!

Addenda

The XO proved itself a profoundly useful edition to our household, particularly for reading e-books. The next generation can be purchased here (Oct. 18, 2008).
Amazon is sold out for now (Jan. 12, 2009).
Brian Stelter reported today in his The New York Times article entitled "Can CNN, the Go-to Site, Get You to Stay?" that cnn.com garners on average 1.7 billion page views per month. The web is the future (01/17/09)!

I could not emphasize the relevance of free global exchange of ideas better than Dr. Goodall in this 2002 presentation (07/07/2010):

Sturmey-Archer

2007-12-06T09:59:00.010-06:00

The problem
My daughter and I used to go to work together at Vanderbilt University in Nashville, Tennessee. The university has got a pretty campus. The trouble was that we both needed to go to places at both ends that were hard to access at rush hour. A huge hospital, Vanderbilt University Medical Center, is nearby and long lines of cars move slowly around every block. Since we do not live very far, we opted for a bike with a back seat for a co-pilot. I am not really a pro biker by any stretch. So we bought a cheap low-riding cruiser and a seat from our local Wal*Mart. Low-riding seemed to be important to keep the center of gravity close to the ground. We had the shortest commute ever and a great time on the way, until it rained ... Boy those caliper breaks don't do very much when they are wet. Though I clamped down on the breaks with all my strength, we slid down the hill almost into traffic and turned away just so before we hit the busy road big time. We both got away with a pretty good fright.

The solution
What's the solution, here? Take the car, whenever there is a droplet in sight? Nope! I thought, it is best to ask people for advice who understand rain: the BRITS. If somebody has got experience with cycling safely drenched, it is got to be the British. They must know a solution to the problem. And certainly, they do! It is the drum break! The break pads are protected from the rain in the sealed hub.

Searching with Google on the net I found Sturmey Archer. Their drum break hubs looked very functional and cool. I found an e-mail order bike shop in the U.S. that carries them (Alfred E Bike) and ordered the X-FD for a front wheel.

The package arrived. The next problem dawned on me right away, when I held the shiny hub in my hand: How do we put this in?
Answer: You have to re-lace the spokes of your wheel. Again, I googled for answers and got them from Sheldon Brown at Harris Cyclery. Though the project seemed fascinating, I decided quickly, it was not going to be me who was going to do this. I did not have the tools, and it seemed to take quite some experience to do it right. I resorted to our local bike store, Cumberland Transit. They did a wonderful job for about $70.-, spokes included. The hub was about $50.-. I managed to install the wheel and connect the actuator to the handle myself. That was not too difficult for someone with average skill.

The picture shows the finished product in use. We had a solution for $120.-, pedaled happily away for another two years rain or shine and took breaks at Harris-Teeter to have Orangina which tastes twice as good from pot-bellied bottles, until my co-pilot grew tall, stepped outside on one chilly winter morning and announced: “It is too cold, Dad. Let's take the car!” Luckily, we had to report to different places then and the traffic was not too bad.

P.S: You may notice, we also installed a Python fork to go with the drum break. It sure makes for a smoother, be it more wobbly ride. If you choose to go down that road, too, make sure that the stem of the fork fits the neck of your frame before you order.

Addendum

Today, I discovered the coolest advertising campaign ever, excelling in simplicity and cleverness. Those of legal beer-drinking age may check out the latest Fat Tire sales effort (05/24/10):

Broken Cups in Your OS X Cupboard

2007-12-05T09:18:00.005-06:00

I work a lot with pictures professionally and have been using Apple computers for most of my career, except my Master's Thesis which I wrote on a DEC PDP 11. I was only a user. However, it always struck me with trepidation that I did not understand a thing of what was going on under the hood. I was eventually roped into computing, when the Apple operating system turned to UNIX around the year 2000. With the introduction of OS X, Apple had opened itself to Open Source computing which opened a new universe to me. Seven years later, I spent way more time with compiling packages than I ought to, and way more often than I liked, I ended up in a real fix. Regardless, the euphoria of empowerment outweighed the disappointment by far.

Today I report on one example of the kinds of adventure one may encounter on this road. Several months ago, I happened on the website of cups printing. Cups is the printing routine used in OS X. I found a pre-compiled binary of a more recent version than I had ready for installation. Since I wanted to be at the cutting edge, I downloaded the offering and installed it without reading the fine print. The result was that I could not print anymore. The fine print said that one had to re-install the operating system, if the update did not work. Oh!

I was not prepared to do that. So I compiled the most recent 1.4.x version for developers from source myself and installed it. After that, I could not even get passed the login window. I had a series of sweaty moments since then. I did not believe the functions that depend on cups! Much of OS X appears to depend on the cups library in one way or another. How can this be in the age decentralization? I managed to fix the problem by re-installing n older version in the single user mode.

Months later, I re-visited the cups website. The pre-compiled packages were gone, and it was recommended to try one the versions 1.3.x. After a number of tries, one of them worked and I offer the result on my SourceForge.net project (Software for Small Budget Science). The package is compiled with Tiger on Intel. It has worked well for me. However, one nuisance remains. The package needs to be reinstalled after each Apple security update. Let my adventure be a warning.

Interests: %	Appreciation: %	Depreciation: %
	years.

[in] :	[cm]:
[lbs]:	[kg] :
	BMI_:

[in] :	[cm]:
[lbs]:	[kg] :