August 19, 2002
Railroad Notes
It is interesting to see that the status of Amtrak and the Acela has remained front page news in Boston and New York. These are not "Amtrak is horrible" or "Amtrak is wonderful" political pieces. These are hard news on "when will Acela be running". The local success of rail is clear from these stories. High speed rail is now an integral piece of local travel, just like the major highways and airports.
The articles also make clear the extent of the travel shift in just one year. Boston to New York the rail managed to gather in 50% vs airline. New York to DC is getting over 65%. The traffic shift and chaos when the high speed was shut down for a few days indicates both the importance of the rail segment, and the continuing management problems at Amtrak. They still have serious problems with keeping people properly informed and running that railroad. But they've done better than earlier years. They do provide information eventually, and they have been very fast in restoring partial service. It took only a few days to repair about half of the trainsets.
On another front, the return to rail for agricultural produce is continuing. The latest quarterlies are that BNSF is carrying at the rate of 85,000 cars/year. These are long haul refrigerated cars that compete with long haul trucks. Some locations they deliver directly to distributors, like New York, and others they meet up with trucks for the local leg. Each rail car is equivalent to between 3 and 4 tractor trailers. The variation is due to most trailers reaching their weight limit before being completely full, while rail cars can reach their volume limit without going over the weight limit. The traffic volume may cross over the 100,000 annual rate by year end.
The US government pretty much killed railroad transport of produce during the 70's, and it has taken a long time for the railroads to make the changes needed to become competitive with trucks. Products need to be delivered on time and at the proper temperature. Dry grain can sit on a siding for days without spoiling. Goods like potatoes cannot. The new rail cars include satellite tracking and monitoring systems so that the location, temperature, air flow, etc. of all cars can be automatically monitored. That limits spoilage from both mistakes and malfunctions. The highly sensitive fresh goods still take the trucks. The produce trains are still slower than trucks by a few days to a week cross country. That makes them suitable for medium priority goods like potatoes, carrots, etc. and low priority produce like frozen goods.
It is interesting to see that the status of Amtrak and the Acela has remained front page news in Boston and New York. These are not "Amtrak is horrible" or "Amtrak is wonderful" political pieces. These are hard news on "when will Acela be running". The local success of rail is clear from these stories. High speed rail is now an integral piece of local travel, just like the major highways and airports.
The articles also make clear the extent of the travel shift in just one year. Boston to New York the rail managed to gather in 50% vs airline. New York to DC is getting over 65%. The traffic shift and chaos when the high speed was shut down for a few days indicates both the importance of the rail segment, and the continuing management problems at Amtrak. They still have serious problems with keeping people properly informed and running that railroad. But they've done better than earlier years. They do provide information eventually, and they have been very fast in restoring partial service. It took only a few days to repair about half of the trainsets.
On another front, the return to rail for agricultural produce is continuing. The latest quarterlies are that BNSF is carrying at the rate of 85,000 cars/year. These are long haul refrigerated cars that compete with long haul trucks. Some locations they deliver directly to distributors, like New York, and others they meet up with trucks for the local leg. Each rail car is equivalent to between 3 and 4 tractor trailers. The variation is due to most trailers reaching their weight limit before being completely full, while rail cars can reach their volume limit without going over the weight limit. The traffic volume may cross over the 100,000 annual rate by year end.
The US government pretty much killed railroad transport of produce during the 70's, and it has taken a long time for the railroads to make the changes needed to become competitive with trucks. Products need to be delivered on time and at the proper temperature. Dry grain can sit on a siding for days without spoiling. Goods like potatoes cannot. The new rail cars include satellite tracking and monitoring systems so that the location, temperature, air flow, etc. of all cars can be automatically monitored. That limits spoilage from both mistakes and malfunctions. The highly sensitive fresh goods still take the trucks. The produce trains are still slower than trucks by a few days to a week cross country. That makes them suitable for medium priority goods like potatoes, carrots, etc. and low priority produce like frozen goods.
rjh at
[link]
August 19, 2002
Morning Coffee Notes
I read Dave Wieners MCN about people developing new stuff and had my own reactions over coffee. These are generalized a bit beyond the question of new stuff that is not patented to the question of where is the really new software.
The problem with the current user interfaces can be summarized in the description: The desktop metaphor only fits work that is normally done on a desktop. The WIMP metaphor has taken off and dominates in the office and software developers' minds because that is work that is normally done at a desktop. But it is a very limited and unexpressive interface.
Compare the WIMP system with a piano keyboard or artist's paintbrush. The WIMP is easier to learn, but hopelessly slow and unexpressive by comparison. The closest it comes is the keyboard. The touch typist has some of the speed and expressive capability of a piano, but it is clearly less capable. Notions like chords, attack, sustain, etc. are basic to piano playing. They are not expressible on a WIMP interface. Similarly, the artist's paintbrush is sensitive to 3D angles, pressure, speed, ink/paint load, etc. Graphics programs struggle through all sorts of complexities in their attempt to provide what is inherent in the brush.
It will take some serious motivation for software developers to grasp this need because they normally work on a desktop. But it is an area where major improvement could happen. Under the present system of patents and copyrights it is also unlikely to emerge. It is complex enough that only the sharing of ideas and the combination of independently developed concepts will lead to success. The barriers imposed by patents and copyrights make such sharing impractical today.
I read Dave Wieners MCN about people developing new stuff and had my own reactions over coffee. These are generalized a bit beyond the question of new stuff that is not patented to the question of where is the really new software.
- Are we looking in the right places? I do not have any specifics to point at, but I have often found myself surprised by some new software concept that emerges from nowhere. These did not arise overnight. They often are not even all that new. I had just been looking in the wrong places.
- Perhaps we are just in the consolidation phase. The software world lives in cycles of Exploration -> Consolidation -> Exploration -> Consolidation -> ....
- Some things are hard. These hard things take lots of
money for people and tools. At the present time this money is only
available from people who demand copyright and patent protection.
Here, there is a significant political content that could be properly
changed.
During the 60's and 70's the primary advanced research organizations all published their results with very few patent restrictions. Bell Labs, Xerox PARC, IBM Research, and ARPA generated huge volumes of freely usable advanced research. At the time, patents were not applicable to mathematics, software, business practices, or other non-tangibles. Copyrights were limited in scope and software licenses were very generous. Even things like the source code for the IBM operating system were easily available to an interested party.
Today, the US government encourages the patenting of all research results and will not fund publicly available research. The Bell system breakup ended the legal structure that made Bell Labs results available to the world. Corporate research bodies now keep most of their work secret.
Things that are hard, such as the creation of Unix and the Internet, require large resources. The creation of Unix was done by a mixture of Bell Labs, Berkeley, and ARPA researchers. I consider Unix creation "complete" with the 1982 release of 4.2bsd. After that there has been enhancement, retreads, and re-creation. But the creation took eight years, involved dozens of dedicated developers plus hundreds of support staff and millions for hardware. The Internet effort was similarly long and hard. We are fortunate that this was back before patents were extended to software and practices. Neither would likely exist if these barriers to innovation had been in place at that time.
The problem with the current user interfaces can be summarized in the description: The desktop metaphor only fits work that is normally done on a desktop. The WIMP metaphor has taken off and dominates in the office and software developers' minds because that is work that is normally done at a desktop. But it is a very limited and unexpressive interface.
Compare the WIMP system with a piano keyboard or artist's paintbrush. The WIMP is easier to learn, but hopelessly slow and unexpressive by comparison. The closest it comes is the keyboard. The touch typist has some of the speed and expressive capability of a piano, but it is clearly less capable. Notions like chords, attack, sustain, etc. are basic to piano playing. They are not expressible on a WIMP interface. Similarly, the artist's paintbrush is sensitive to 3D angles, pressure, speed, ink/paint load, etc. Graphics programs struggle through all sorts of complexities in their attempt to provide what is inherent in the brush.
It will take some serious motivation for software developers to grasp this need because they normally work on a desktop. But it is an area where major improvement could happen. Under the present system of patents and copyrights it is also unlikely to emerge. It is complex enough that only the sharing of ideas and the combination of independently developed concepts will lead to success. The barriers imposed by patents and copyrights make such sharing impractical today.
rjh at
[link]
August 16, 2002
Discussing the Unthinkable
The bureaucrats and media seem unable to deal with preparing the public for possible terrorist attacks with weapons like nerve gas. The media seems fixated on hype and hysteria. The nonsense around West Nile virus is typical. They take a disease which is only a little more dangerous than the flu and turn it into a national nightmare. The Homeland Security people feed the hype and hysteria over gas because public safety takes a distant third to bureaucratic empire building and politics in their priorities. In fact, hype and hysteria are the greatest danger when dealing with nerve gas.
Nerve gas has been involved in 15 terrorist attacks since 1975. Two of these were large, involved significant casualties, and are well documented. These are the attacks in Matsumoto and Tokyo Japan. These are widely mentioned in the press and by the bureaucrats, but they are very inaccurately described. These errors feed the hype and hysteria that will cause real harm in the event of a real attack.
The attack in Matsumoto was in a residential area at 10:45PM on 27 June 1994:
The Tokyo attack was at 8:00 AM (peak rush hour), 20 March 1995:
The human body recovers from nerve gas very quickly. Once the exposure ends, rest and fresh air are very effective curatives. Most treatment tries to prevent death or injury during the period while the nervous system is disabled.
Over 90% of those exposed and with symptoms in Matsumoto were fully recovered within 24 hours. In Tokyo, over 80% of those hospitalized with gas symptoms were recovered and released within 24 hours. The press releases talk about "5,000 casualties" in Tokyo, but they do not mention that over 80% were panic related and that less than 200 people were killed or hospitalized overnight for gas exposure.
Several important conclusions result from these experiences:
I hope that eventually responsible persons will replace the current bureaucrats and sensible guidance widely issued. In the meantime, a few simple guidelines:
The bureaucrats and media seem unable to deal with preparing the public for possible terrorist attacks with weapons like nerve gas. The media seems fixated on hype and hysteria. The nonsense around West Nile virus is typical. They take a disease which is only a little more dangerous than the flu and turn it into a national nightmare. The Homeland Security people feed the hype and hysteria over gas because public safety takes a distant third to bureaucratic empire building and politics in their priorities. In fact, hype and hysteria are the greatest danger when dealing with nerve gas.
Nerve gas has been involved in 15 terrorist attacks since 1975. Two of these were large, involved significant casualties, and are well documented. These are the attacks in Matsumoto and Tokyo Japan. These are widely mentioned in the press and by the bureaucrats, but they are very inaccurately described. These errors feed the hype and hysteria that will cause real harm in the event of a real attack.
The attack in Matsumoto was in a residential area at 10:45PM on 27 June 1994:
| Attack |
20 kg Sarin, Fan dispersed from a stationary
van on a calm night, Peak toxicity area 400m x 300m ellipse |
| Casualties |
7 dead 51 hospitalized 253 treated and released 277 exposed with symptoms, untreated. (These were not known until subsequent investigation.) |
The Tokyo attack was at 8:00 AM (peak rush hour), 20 March 1995:
| Attack |
4.5 kg Sarin total. Released in 5 subway cars
on 5 different routes by spilling sarin onto the floor of the subway cars.
The intent was to spill 7kg, but some of the packages malfunctioned. |
| Casualties |
12 dead 17 critical illness 37 severe illness 984 moderate illness 4,470 hospitalized for panic and hysteria illnesses. Not exposed. |
The human body recovers from nerve gas very quickly. Once the exposure ends, rest and fresh air are very effective curatives. Most treatment tries to prevent death or injury during the period while the nervous system is disabled.
Over 90% of those exposed and with symptoms in Matsumoto were fully recovered within 24 hours. In Tokyo, over 80% of those hospitalized with gas symptoms were recovered and released within 24 hours. The press releases talk about "5,000 casualties" in Tokyo, but they do not mention that over 80% were panic related and that less than 200 people were killed or hospitalized overnight for gas exposure.
Several important conclusions result from these experiences:
- Panic and hysteria are a major problem, and quite unnecessary. The burden of dealing with the panic patients was a serious problem in Tokyo. It interfered with treating those who were actually exposed.
- Prompt evacuation worked.
- Procedures developed for chemical spills worked.
- Fatal modesty will be a problem. Some of the exposures were caused by people wearing contaminated clothes into the hospital and exposing patients and staff.
- Nerve gas is only somewhat more deadly per pound than human bombs, car bombs, and other explosive devices. This is because it is hard to use. (This is also consistent with Pentagon estimates of 10,000 casualties per ton of nerve gas for unprotected troops in open terrain.)
I hope that eventually responsible persons will replace the current bureaucrats and sensible guidance widely issued. In the meantime, a few simple guidelines:
- If attacked, leave the area fast.
- Follow the instructions of police and fire. The existing procedures for chemical spill evacuations are appropriate for nerve gas.
- Follow instructions for decontamination. If this means showing the public that embarassing underwear, do it. If it means public nakedness, do it. You are risking your life and threatening the lives of many other people if you do not decontaminate.
- Don't panic or worry about minor symptoms after you leave the exposure area. If you are conscious, breathing, and able to control your body, then you will recover. In a sense, if you are healthy enough to worry, then you are healthy enough that clean air will cure you.
rjh at
[link]
August 15, 2002
Wind Power on the way to P-town
I took the ferry to P-town for lunch yesterday. It was a good day to skip work. It probably hit 40C downtown. It was 38C at Logan. The day was also a wind power show.
First, we had Boston's new modern wind mill. It's on Windmill Point, and as you might guess from the name, it's not the first one there. I think that there have been five or six windmills there since colonial times. This one is a new 660KWp model. It was gently turning in the breeze. It was hard to judge the winds while on the ferry, but later examination on land indicated 10-15 kt breeze all day.
Then in addition to the usual little sloops, we got two three-masted barques. The first was a mystery ship off on the horizon. It was well west of us, and at one point it looked like shimmering sails hovering above the ocean haze in the morning sun. At this long range we could not identify it. All that could be told was that it was a three masted barque. Later web searching reveals that it is the replica of the barque HMS Bounty. (Yes, the famous one.)
Further east we found the USCG barque Eagle gently sliding toward Boston under sail.
Finally in Ptown there was the usual collection of panels and pinwheels on many of the larger sloops and ketches.
All this points to the wind power resources along the coastline and in the near coastal waters off Massachusetts. The Boston Globe has recently recognized the commercial viability of wind power in its own unique manner. They realized that someone was going to make money, so they editorialized that it is time to regulate and tax it.
The total production from the two farms being planned would be about 1,300 MWp. The shoals could easily support 3,000 MWp before running into limits imposed by tourism, shipping, fishing, wildlife, and national parks. These sites would just be the shoals that are difficult to see from shore and avoid any of those other uses. Installing 3,000 MWp would bring production into the range that exposes the next barrier to the use of wind energy in the United States. The long distance power grid is designed only for electricity producers that can start and stop production on a schedule.
This has been less of a problem in Europe because their many wind power systems are also located relatively closely to the electricity consumers. The distances involved are often less than 100km. They have only faced the problems and costs of extending higher capacity power lines to the scattered local wind farms. They have not had to deal with the complexities of power wheeling. The PV generators and smaller wind systems in the US have also been able to avoid the grid problem due to their small size.
The long distance grid has many choke points and capacity limits. It depends upon generators and users scheduling their power shipments accurately. To motivate compliance with these schedules, the grid operators charge steep penalties for missing a schedule. These can be 2-3 cents per kilowatt hour. Penalties like this are a serious problem for the wind energy producers. The winds are hard to schedule and these penalties make it impossible to ship power profitably.
The grid system will need both physical and organizational changes. It needs design changes to reduce the vulnerability to missed schedules and to permit short term schedule changes. It needs organizational changes to separate penalties for mismanagement and "gaming" the system from the real costs of missing schedules. This kind of change takes years, so the efforts are just beginning now.
It may be a little bit premature, but the chokepoint and capacity problems will take a long time to solve and affect much more than wind power. Most of them derive from the conflict between local needs and regional or national needs. For example, there is a major political conflict between CT and NY over a power line under Long Island sound. Why should CT have all that environmental impact and potential power problems to help NY, especially when part of the problem is caused by Long Islands refusal to accept the environmental impact of building power generators? There are dozens of similar battles around the country where local interests are hurt by grid improvements that are of no local value.
These first two wind farms will not face this problem immediately. At full power they will meet the loads of Cape Cod and some of southeast MA. So the long distance grids don't need to be dramatically upgraded to carry the wind power elsewhere. When these farms are running, the power that would have gone to Cape Cod can be shipped elsewhere. The local power grid needs some local improvements to add high capacity circuits to the landing points for these farms, but that has a corresponding local benefit.
Wind power has other problems when it becomes a significant fraction of the local power supply, but those will be discussed at another time.
I took the ferry to P-town for lunch yesterday. It was a good day to skip work. It probably hit 40C downtown. It was 38C at Logan. The day was also a wind power show.
First, we had Boston's new modern wind mill. It's on Windmill Point, and as you might guess from the name, it's not the first one there. I think that there have been five or six windmills there since colonial times. This one is a new 660KWp model. It was gently turning in the breeze. It was hard to judge the winds while on the ferry, but later examination on land indicated 10-15 kt breeze all day.
Then in addition to the usual little sloops, we got two three-masted barques. The first was a mystery ship off on the horizon. It was well west of us, and at one point it looked like shimmering sails hovering above the ocean haze in the morning sun. At this long range we could not identify it. All that could be told was that it was a three masted barque. Later web searching reveals that it is the replica of the barque HMS Bounty. (Yes, the famous one.)
Further east we found the USCG barque Eagle gently sliding toward Boston under sail.
Finally in Ptown there was the usual collection of panels and pinwheels on many of the larger sloops and ketches.
All this points to the wind power resources along the coastline and in the near coastal waters off Massachusetts. The Boston Globe has recently recognized the commercial viability of wind power in its own unique manner. They realized that someone was going to make money, so they editorialized that it is time to regulate and tax it.
The total production from the two farms being planned would be about 1,300 MWp. The shoals could easily support 3,000 MWp before running into limits imposed by tourism, shipping, fishing, wildlife, and national parks. These sites would just be the shoals that are difficult to see from shore and avoid any of those other uses. Installing 3,000 MWp would bring production into the range that exposes the next barrier to the use of wind energy in the United States. The long distance power grid is designed only for electricity producers that can start and stop production on a schedule.
This has been less of a problem in Europe because their many wind power systems are also located relatively closely to the electricity consumers. The distances involved are often less than 100km. They have only faced the problems and costs of extending higher capacity power lines to the scattered local wind farms. They have not had to deal with the complexities of power wheeling. The PV generators and smaller wind systems in the US have also been able to avoid the grid problem due to their small size.
The long distance grid has many choke points and capacity limits. It depends upon generators and users scheduling their power shipments accurately. To motivate compliance with these schedules, the grid operators charge steep penalties for missing a schedule. These can be 2-3 cents per kilowatt hour. Penalties like this are a serious problem for the wind energy producers. The winds are hard to schedule and these penalties make it impossible to ship power profitably.
The grid system will need both physical and organizational changes. It needs design changes to reduce the vulnerability to missed schedules and to permit short term schedule changes. It needs organizational changes to separate penalties for mismanagement and "gaming" the system from the real costs of missing schedules. This kind of change takes years, so the efforts are just beginning now.
It may be a little bit premature, but the chokepoint and capacity problems will take a long time to solve and affect much more than wind power. Most of them derive from the conflict between local needs and regional or national needs. For example, there is a major political conflict between CT and NY over a power line under Long Island sound. Why should CT have all that environmental impact and potential power problems to help NY, especially when part of the problem is caused by Long Islands refusal to accept the environmental impact of building power generators? There are dozens of similar battles around the country where local interests are hurt by grid improvements that are of no local value.
These first two wind farms will not face this problem immediately. At full power they will meet the loads of Cape Cod and some of southeast MA. So the long distance grids don't need to be dramatically upgraded to carry the wind power elsewhere. When these farms are running, the power that would have gone to Cape Cod can be shipped elsewhere. The local power grid needs some local improvements to add high capacity circuits to the landing points for these farms, but that has a corresponding local benefit.
Wind power has other problems when it becomes a significant fraction of the local power supply, but those will be discussed at another time.
rjh at
[link]
August 13, 2002
Solar Power in Kenya
The report on Amorphous PV in Kenya is an example of how the market is gradually shifting into a much more sensible engineering and accounting driven world and away from the feel good emotional world. It discusses the basic market for commercially successful PV in Kenya. The limitations of the Kenyan grid and the typical needs of the Kenyan customer are summarized. Most buyers are rural users who construct a home-brew solar system out of a small PV panel (10 to 15W) and some batteries. This will power some lights and small electronics (TV, radio) during the evening. It is unlikely that grid power will reach these locations for many years. These small solar systems are a major home improvement investment. They typically reflect 4-6 months income.
There is a reliability analysis that is driven by concerns arising from reported problems with the PV modules. They found significant problems with one vendor. They also note local issues regarding warranty repairs. The vendors honor their warranty. The problem is that with the home brew systems it is often hard to determine whether it is the PV panel, the battery, or something else that has a problem. Also, there is typically no warning of problems before the system simply fails. The inability to diagnose or predict problems leads to extra costs.
The rest of the analysis is that the panels do not deliver the advertised power but that this is an endemic problem with all panels from all vendors. The advertised power is 10-20% higher than the delivered power.
Overall, these small PV systems are a commercial success and do satisfy a real customer need. There is some marketing exaggeration, but not major fraud. There is one vendor with serious quality problems, but the others are doing well. Sales range between 10,000 to 20,000 units per year, and are growing. There is no local price subsidy for the units. The ups and downs of the local economy are a major factor in sales fluctuations. It is locally a profitable commercial activity, and is gradually spreading from Kenya to neighboring countries.
They do not discuss the profitability of the panel vendors. I checked the web site of Free Energy Europe, which was the lowest price of the vendors examined. They are a Dutch firm but do not have public financials. The problem vendor Intersolar has public financials. It is presently losing money, has enough cash to run for two years at that rate, has recently increased production capacity, and has increasing sales. They look like the typical risky small company startup.
Correction below on Dye PV
I had the power generation numbers wrong for that overhang. It would not be at the low limit of reasonable. It would be an efficient size for inverters. Corrected below. I should also note that the DC-AC inverter for such a system is likely to cost about $10K installed. The cost of ancillary components like inverters becomes a significant item as PV costs drop.
Economics of HE washers
I complained earlier about the questionable accounting from some eco-freaks who claimed a 45% ROI for high efficiency washing machines. So I did a little spot analysis using my own numbers. In my case I see no saving from reduced water usage. I was at the minimum bill before, and I'm at the minimum bill now. I've always been below the basic use rate. I did see an electricity saving. The amounts bounce around with other variables, but it looked like a saving of about $3/month. This is consistent with the range of savings expected. The saving is primarily from superior water extraction that reduces the drying load, and somewhat from reduced hot water usage. With gas for hot water and drying, the cost savings are typically half that of electricity.
I then examined current pricing. There are two pricing factors. First, the high efficiency washers are only available in fancy models with lots of features. There is no minimal washer option. Second, for equivalent feature set the high efficiency models cost about $250 more. This is not a fancy exhaustive study. This was visiting a couple stores and checking prices.
So I then computed the IRR for several combinations of costs and savings. I get about $36 per year savings in a somewhat above average electricity cost area. So I computed for savings of $45, $30, and $15 per year to cover the range of electrical and gas savings. Then there are sometimes discounts that reduce the cost increment. At times there have been $100 and $200 discounts locally. So I assumed an incremental investment of $250 and $150. Then I computed the IRR:
In my case, I got a decent IRR. Assuming 10 year life I did better than the bond rate; for a 20 year life it matches the long term stock market rate. So HE washers will make sense for people who want the fancier washers and who have high local costs. If I had gas in the house, then it would be a loser.
These numbers are also much more believable than that 42% ROI nonsense from the eco-freaks. It says that if you will get a good lifespan (10+ years), a good savings (30/yr or more), can get one of those discount deals to reduce the cost difference, and want a fancier washer, then buying the high efficiency is a good investment. Absent this, you need either a long time horizon or larger cost savings. You might still want one of these for some other reason, but it is not a good ecology investment. Odds are that there is some other investment that you could make which would accomplish similar ecological savings and have a better IRR for you.
This analysis ignores inflation effects.
The report on Amorphous PV in Kenya is an example of how the market is gradually shifting into a much more sensible engineering and accounting driven world and away from the feel good emotional world. It discusses the basic market for commercially successful PV in Kenya. The limitations of the Kenyan grid and the typical needs of the Kenyan customer are summarized. Most buyers are rural users who construct a home-brew solar system out of a small PV panel (10 to 15W) and some batteries. This will power some lights and small electronics (TV, radio) during the evening. It is unlikely that grid power will reach these locations for many years. These small solar systems are a major home improvement investment. They typically reflect 4-6 months income.
There is a reliability analysis that is driven by concerns arising from reported problems with the PV modules. They found significant problems with one vendor. They also note local issues regarding warranty repairs. The vendors honor their warranty. The problem is that with the home brew systems it is often hard to determine whether it is the PV panel, the battery, or something else that has a problem. Also, there is typically no warning of problems before the system simply fails. The inability to diagnose or predict problems leads to extra costs.
The rest of the analysis is that the panels do not deliver the advertised power but that this is an endemic problem with all panels from all vendors. The advertised power is 10-20% higher than the delivered power.
Overall, these small PV systems are a commercial success and do satisfy a real customer need. There is some marketing exaggeration, but not major fraud. There is one vendor with serious quality problems, but the others are doing well. Sales range between 10,000 to 20,000 units per year, and are growing. There is no local price subsidy for the units. The ups and downs of the local economy are a major factor in sales fluctuations. It is locally a profitable commercial activity, and is gradually spreading from Kenya to neighboring countries.
They do not discuss the profitability of the panel vendors. I checked the web site of Free Energy Europe, which was the lowest price of the vendors examined. They are a Dutch firm but do not have public financials. The problem vendor Intersolar has public financials. It is presently losing money, has enough cash to run for two years at that rate, has recently increased production capacity, and has increasing sales. They look like the typical risky small company startup.
Correction below on Dye PV
I had the power generation numbers wrong for that overhang. It would not be at the low limit of reasonable. It would be an efficient size for inverters. Corrected below. I should also note that the DC-AC inverter for such a system is likely to cost about $10K installed. The cost of ancillary components like inverters becomes a significant item as PV costs drop.
Economics of HE washers
I complained earlier about the questionable accounting from some eco-freaks who claimed a 45% ROI for high efficiency washing machines. So I did a little spot analysis using my own numbers. In my case I see no saving from reduced water usage. I was at the minimum bill before, and I'm at the minimum bill now. I've always been below the basic use rate. I did see an electricity saving. The amounts bounce around with other variables, but it looked like a saving of about $3/month. This is consistent with the range of savings expected. The saving is primarily from superior water extraction that reduces the drying load, and somewhat from reduced hot water usage. With gas for hot water and drying, the cost savings are typically half that of electricity.
I then examined current pricing. There are two pricing factors. First, the high efficiency washers are only available in fancy models with lots of features. There is no minimal washer option. Second, for equivalent feature set the high efficiency models cost about $250 more. This is not a fancy exhaustive study. This was visiting a couple stores and checking prices.
So I then computed the IRR for several combinations of costs and savings. I get about $36 per year savings in a somewhat above average electricity cost area. So I computed for savings of $45, $30, and $15 per year to cover the range of electrical and gas savings. Then there are sometimes discounts that reduce the cost increment. At times there have been $100 and $200 discounts locally. So I assumed an incremental investment of $250 and $150. Then I computed the IRR:
| Incremental cost, Annual Saving |
250, 45 |
250, 30 |
250, 15 |
150, 30 |
150, 15 |
| Five year life |
-3% |
-15% |
-30% |
0 |
-19% |
| Ten year life |
12 |
3 |
-8 |
15 |
0 |
| Twenty year life |
17 |
10 |
2 |
19 |
8 |
In my case, I got a decent IRR. Assuming 10 year life I did better than the bond rate; for a 20 year life it matches the long term stock market rate. So HE washers will make sense for people who want the fancier washers and who have high local costs. If I had gas in the house, then it would be a loser.
These numbers are also much more believable than that 42% ROI nonsense from the eco-freaks. It says that if you will get a good lifespan (10+ years), a good savings (30/yr or more), can get one of those discount deals to reduce the cost difference, and want a fancier washer, then buying the high efficiency is a good investment. Absent this, you need either a long time horizon or larger cost savings. You might still want one of these for some other reason, but it is not a good ecology investment. Odds are that there is some other investment that you could make which would accomplish similar ecological savings and have a better IRR for you.
This analysis ignores inflation effects.
rjh at
[link]
August 11, 2002
OODA and Medical Practice
This was originally going to be a writeup on papers from Decision Support Systems, but they were disappointing. They start with an interesting simple premise:
First, for those who have forgotten, OODA is for Orient, Observe, Decide, Act. This terminology began with the Air Force analysis of aerial combat, but it applies to many kinds of scientific, commercial, and other activities. The use of "orient" is sometimes confusing. It means to gather context information to decide what observations are needed to make the decisions for this turn of the cycle. It is fairly easy to see how OODA applies to much of medical practice.
There is little (if any) systematic optimization of the OODA cycle in medicine. Medical training teaches specific tools and methods that are used during the OODA cycle, but does not spend much time on the cycle itself. Experience provides a qualitative feel for the cycle, but it does not provide the understanding needed to redesign the cycle. Some of the basic questions that are rarely answered are:
There is little effort at present to directly examine and understand the medical OODA cycle. Techniques like evidence based medicine are a start, and may lead partially in that direction. Administrative workflow analysis often improves OODA behavior because it often reveals obviously unproductive bottlenecks and delays. There are many people working to improve diagnostic tools and make things faster. But these are targeting components, not the entire cycle. Without the understanding of the full cycle you cannot focus efforts where they will have the greatest impact.
Conscious analysis of the OODA cycle in other disciplines has lead to significant improvements. Medicine needs to do this also.
This was originally going to be a writeup on papers from Decision Support Systems, but they were disappointing. They start with an interesting simple premise:
- The practice of medicine is an OODA system
- Improvement requires conscious OODA optimization
First, for those who have forgotten, OODA is for Orient, Observe, Decide, Act. This terminology began with the Air Force analysis of aerial combat, but it applies to many kinds of scientific, commercial, and other activities. The use of "orient" is sometimes confusing. It means to gather context information to decide what observations are needed to make the decisions for this turn of the cycle. It is fairly easy to see how OODA applies to much of medical practice.
There is little (if any) systematic optimization of the OODA cycle in medicine. Medical training teaches specific tools and methods that are used during the OODA cycle, but does not spend much time on the cycle itself. Experience provides a qualitative feel for the cycle, but it does not provide the understanding needed to redesign the cycle. Some of the basic questions that are rarely answered are:
- What is the optimum pace for the OODA cycle for various illnesses and situations?
- How should administrative procedures be organized to enhance OODA performance
- How are OODA cycle components limited by disease processes; human decision processes; technology processes; administrative process?
There is little effort at present to directly examine and understand the medical OODA cycle. Techniques like evidence based medicine are a start, and may lead partially in that direction. Administrative workflow analysis often improves OODA behavior because it often reveals obviously unproductive bottlenecks and delays. There are many people working to improve diagnostic tools and make things faster. But these are targeting components, not the entire cycle. Without the understanding of the full cycle you cannot focus efforts where they will have the greatest impact.
Conscious analysis of the OODA cycle in other disciplines has lead to significant improvements. Medicine needs to do this also.
rjh at 2300
[link]
August 11, 2002
Russian English orthography
The front page of Sudarushka has a big picture of Hallie Barrie and they transliterate the "H" in her name as a cha! (Looks like an X). But they use the normal convention and translate the H in "Hollywood" as a "ge", (looks like a gamma). Does this mean that the Russians are starting to shift and translate the sound match better now that they have much more interaction with English?
Who knows. You can't read too much into the practices of a celebrity magazine (analagous to "Us" or "People"). I was just very surprised when I saw that front page.
More electronics headaches
A bad day for electronics here. My backup computer died. Well at least partly dead. It no longer speaks ethernet to anyone. Its power supply fan makes noises. These are both bad behavior. It is called the backup computer because that's what it does. It keeps backup copies of stuff on my network. So it has no CRT or keyboard to tell me anything about what is really going wrong. It could be a bad disk drive too. Surgery will be needed. The thing was mostly made from salvage parts, so it may just be time to tear it apart and rebuilt it again.
And the little room fan that circulates air for the computers died.
I know I'd been getting whiffs of electronics failure the past couple days. Was it the computer, the fan, or something else that hasn't broken yet.
The front page of Sudarushka has a big picture of Hallie Barrie and they transliterate the "H" in her name as a cha! (Looks like an X). But they use the normal convention and translate the H in "Hollywood" as a "ge", (looks like a gamma). Does this mean that the Russians are starting to shift and translate the sound match better now that they have much more interaction with English?
Who knows. You can't read too much into the practices of a celebrity magazine (analagous to "Us" or "People"). I was just very surprised when I saw that front page.
More electronics headaches
A bad day for electronics here. My backup computer died. Well at least partly dead. It no longer speaks ethernet to anyone. Its power supply fan makes noises. These are both bad behavior. It is called the backup computer because that's what it does. It keeps backup copies of stuff on my network. So it has no CRT or keyboard to tell me anything about what is really going wrong. It could be a bad disk drive too. Surgery will be needed. The thing was mostly made from salvage parts, so it may just be time to tear it apart and rebuilt it again.
And the little room fan that circulates air for the computers died.
I know I'd been getting whiffs of electronics failure the past couple days. Was it the computer, the fan, or something else that hasn't broken yet.
Another PV technology
I drove past a hotel today and noticed their large protective overhang. It protected the driveways and blended up into the windows for an entry and reception atrium. The whole thing faces south for good sunlight. It won't be too long until construction like that will be photovoltaic PV power generation.
The next new thing in PV are the dye based PV. This "new" technology was covered in laboratory reports and experiments during 1990-1995. The laboratory work continues with the goal of increasing its efficiency and reducing costs. The early production is now starting to be sold and installed. It will probably get a burst of publicity in conjunction with an Australian conference and the transition to commercial production for the Australian manufacturer.
The underlying concept is the use of a dye that is sandwiched between semi-conductors and conducting layers. The dye absorbs the sunlight, which is then converted to electricity. The important novelty is that this approach does not use silicon or high temperature vacuum deposition manufacturing. These have the potential for significant reduction in manufacturing costs. The problem with dye PV is the low efficiency at present. Crystalline silicon PV gets 15% efficiency, amorphous silicon 10%, and dye PV only 5% efficiency. So you need three times the area for the same power with dye PV. But the manufacturing cost per Watt is still much lower.
The initial push for dye PV will be locations that were already going to be built, such as that hotel overhang. You don't get a good payback from tripling the size of additional construction, but adding a little extra to an already planned component can have a good ROI. If the extra cost for the dye PV component is $2 per Wp (peak Watt in full lighting) you do find suitable locations. Most of the US gets about 2000 equivalent hours of full sun. That overhang was already angled for decent sun orientation and the dye PV tolerate off angle lighting of 20 degrees without much loss. So for each extra dollar they would generate about 1KWh per year.
There will also be some size related considerations. That overhang was about 15m x 20m, so at 5% efficiency it would be generating up to1500W10.5 kW. That is near the small limit
for cost effective conversion to AC. That is a reasonably
efficient size for power conversion to AC. You need an inverter and
safety approved electrical hookup. These all add to the cost. It
also helps to have at least a rudimentary internal monitoring system
so that you can keep enough records to show the return on the
investment. The inverter makers are learning more about their
customers and now offer simple record keeping as part of the
inverter.
The ROI analysis then depends upon the local electric structure. Peak load adjustable rates can make the effective rate during peak sun periods $0.15/KWh or more. More remote locations, like islands, may be significantly higher. If you pick the locations carefully there will be sites that make a positive return. You need good sun and very high local electrical rates.
The initial production from the Australian firm is going to be window wall components. When laminated onto window glass, the dye PV acts much like the tinting layers used to cut sun glare. It also provides insulating enhancement. Most office buildings have significant expanses of south facing windows, and many have areas of angled window with good sun orientation. There is probably enough such construction scattered around the Australian area to keep the one factory busy and to build an experience base. With time the experience will lead to lower cost manufacturing and expand the potential market to a larger portion of the building industry.
The dye PVs will be in competition with the Amorphous PV. Amorphous is a much older technology that has recently had a growth spurt. A 30 MWp/yr factory went online this year in Michigan. It coats silicon onto stainless steel to create rugged flexible PV. The efficiency characteristics of amorphous remain a sales problem. It comes out of the factory almost as good as crystalline, then plummets to about 10% efficiency in the first six months. After that the efficiency gradually drifts down. It can be restored by removal and heat treatment, but that is not practical in the field. It mostly makes recycling of old used panels practical. The current effort has been in slowing the degradation so that panels retain a 10+ year useful life. After that, they may be removed for heat treatment and resale. That hotel could have used dye for both the overhang and the windows. Or it could have used Amorphous for the overhang. The cost and performance tradeoffs will remain complex unless the dye researchers find a way to substantially improve its efficiency.
On a vaguely related matter, I checked out the market numbers for high efficiency washing machines. These have been slowly picking up market share. Right now they are at 3-5% of washing machines sold. In 1995, about 1% and 1990 about 0.5%. I dug those numbers out because I got to wondering why my local stores did not carry the HE detergent that I need. I had to wander over to a Walmart to pick it up. This said something about market penetration. If I assume a typical 20-yr life for washing machines (taking into consideration resale as used, etc.) the answer jumps out. HE machines are only 1-2% of the installed base.
The relationship? That hotel was on the way to Walmart. And yes, the detergent does matter. Regular detergent foams up the machine really badly, so you either don't use enough or have foam problems. HE machines don't need the foam effect because the tumble wash provides the aeration. Top loading agitators need the foaming agents to get the cleaning effect.
And while at it I found more dishonest advertising by the eco activists. They were claiming a 45% ROI for high efficiency machines. This sounded absurd to me. With some digging through their numbers I figured out that what they really meant was that there was a 45% ROI on the difference between ordinary and HE machines. But they did not give enough additional information to check their cost assumptions. Given the dishonesty of their summary, I doubt the accuracy of their accounting.
I wonder if they realize how much they hurt their own case with their dishonest accounting summaries.
I drove past a hotel today and noticed their large protective overhang. It protected the driveways and blended up into the windows for an entry and reception atrium. The whole thing faces south for good sunlight. It won't be too long until construction like that will be photovoltaic PV power generation.
The next new thing in PV are the dye based PV. This "new" technology was covered in laboratory reports and experiments during 1990-1995. The laboratory work continues with the goal of increasing its efficiency and reducing costs. The early production is now starting to be sold and installed. It will probably get a burst of publicity in conjunction with an Australian conference and the transition to commercial production for the Australian manufacturer.
The underlying concept is the use of a dye that is sandwiched between semi-conductors and conducting layers. The dye absorbs the sunlight, which is then converted to electricity. The important novelty is that this approach does not use silicon or high temperature vacuum deposition manufacturing. These have the potential for significant reduction in manufacturing costs. The problem with dye PV is the low efficiency at present. Crystalline silicon PV gets 15% efficiency, amorphous silicon 10%, and dye PV only 5% efficiency. So you need three times the area for the same power with dye PV. But the manufacturing cost per Watt is still much lower.
The initial push for dye PV will be locations that were already going to be built, such as that hotel overhang. You don't get a good payback from tripling the size of additional construction, but adding a little extra to an already planned component can have a good ROI. If the extra cost for the dye PV component is $2 per Wp (peak Watt in full lighting) you do find suitable locations. Most of the US gets about 2000 equivalent hours of full sun. That overhang was already angled for decent sun orientation and the dye PV tolerate off angle lighting of 20 degrees without much loss. So for each extra dollar they would generate about 1KWh per year.
There will also be some size related considerations. That overhang was about 15m x 20m, so at 5% efficiency it would be generating up to
The ROI analysis then depends upon the local electric structure. Peak load adjustable rates can make the effective rate during peak sun periods $0.15/KWh or more. More remote locations, like islands, may be significantly higher. If you pick the locations carefully there will be sites that make a positive return. You need good sun and very high local electrical rates.
The initial production from the Australian firm is going to be window wall components. When laminated onto window glass, the dye PV acts much like the tinting layers used to cut sun glare. It also provides insulating enhancement. Most office buildings have significant expanses of south facing windows, and many have areas of angled window with good sun orientation. There is probably enough such construction scattered around the Australian area to keep the one factory busy and to build an experience base. With time the experience will lead to lower cost manufacturing and expand the potential market to a larger portion of the building industry.
The dye PVs will be in competition with the Amorphous PV. Amorphous is a much older technology that has recently had a growth spurt. A 30 MWp/yr factory went online this year in Michigan. It coats silicon onto stainless steel to create rugged flexible PV. The efficiency characteristics of amorphous remain a sales problem. It comes out of the factory almost as good as crystalline, then plummets to about 10% efficiency in the first six months. After that the efficiency gradually drifts down. It can be restored by removal and heat treatment, but that is not practical in the field. It mostly makes recycling of old used panels practical. The current effort has been in slowing the degradation so that panels retain a 10+ year useful life. After that, they may be removed for heat treatment and resale. That hotel could have used dye for both the overhang and the windows. Or it could have used Amorphous for the overhang. The cost and performance tradeoffs will remain complex unless the dye researchers find a way to substantially improve its efficiency.
On a vaguely related matter, I checked out the market numbers for high efficiency washing machines. These have been slowly picking up market share. Right now they are at 3-5% of washing machines sold. In 1995, about 1% and 1990 about 0.5%. I dug those numbers out because I got to wondering why my local stores did not carry the HE detergent that I need. I had to wander over to a Walmart to pick it up. This said something about market penetration. If I assume a typical 20-yr life for washing machines (taking into consideration resale as used, etc.) the answer jumps out. HE machines are only 1-2% of the installed base.
The relationship? That hotel was on the way to Walmart. And yes, the detergent does matter. Regular detergent foams up the machine really badly, so you either don't use enough or have foam problems. HE machines don't need the foam effect because the tumble wash provides the aeration. Top loading agitators need the foaming agents to get the cleaning effect.
And while at it I found more dishonest advertising by the eco activists. They were claiming a 45% ROI for high efficiency machines. This sounded absurd to me. With some digging through their numbers I figured out that what they really meant was that there was a 45% ROI on the difference between ordinary and HE machines. But they did not give enough additional information to check their cost assumptions. Given the dishonesty of their summary, I doubt the accuracy of their accounting.
I wonder if they realize how much they hurt their own case with their dishonest accounting summaries.
rjh at 2300
[link]
August 5, 2002
Copyright and Free Downloads
Some very interesting articles related to how free downloads interact with the market for copyrighted materials. Janis Ian and Courney Love present their perspectives. From Baen Books you can download portions (and sometimes complete) books. Their experience has been that this kind of structured downloading has significantly boosted sales of paper books. The National Academy Press has experienced a similar increase in sales when they made all of their books available for download. Both discuss their experiences.
This is somewhat old stuff, but I figured that it is worth capturing these links in one place. Only the Janis Ian article is really new.
Some very interesting articles related to how free downloads interact with the market for copyrighted materials. Janis Ian and Courney Love present their perspectives. From Baen Books you can download portions (and sometimes complete) books. Their experience has been that this kind of structured downloading has significantly boosted sales of paper books. The National Academy Press has experienced a similar increase in sales when they made all of their books available for download. Both discuss their experiences.
This is somewhat old stuff, but I figured that it is worth capturing these links in one place. Only the Janis Ian article is really new.
rjh at 2300
[link]
Laptop Lust
For a while I've been lusting after a laptop. This doesn't mean that I will actually go buy one. Half the fun is doing the research, evaluations, and shopping. After all that I may still decide to wait another six months or a year. Then I can do it all over again.
I don't need a laptop. The desktop network has been just fine. But it would be nice to be able to roam around the house instead of being stuck at a desk. Most of my personal work is Unix based, so the laptop choice is more complex. It would be primarily operating as an X-server communicating over SSH encrypted links (over WEP over 802.11b) to the real computer. That continues my practice of preserving the same user view regardless of what computer I am using. It also deals with the security risks of wireless work. So what I need is not CPU power, but X-windows and Unix capability. As a secondary consideration I might take a laptop traveling, but it would primarily be for network access, so the requirements are similar.
I've had corporate laptops for over 10 years. The first one was a 386 laptop, and I've had some kind of laptop ever since. I could use my current employer's laptop, but I practice extremely strict hardware separation of personal from employer use. I think that this is appropriate from an ethical, moral, legal, and practical perspective. That last may need some explanation. I've experienced and witnessed thefts, takeovers, and bankruptcies. There is a very real chance that I might lose the laptop without warning. It would be just suddenly gone and in the hands of someone who might try to use its contents against me. I structure my computer use such that the sudden loss of the corporate laptop would not disturb my personal life.
So I have enough experience to confidently define my ideal laptop:
Then on to the PC world. Lots of fun looking through web sites. I ended up investigating the ones offered by my favorite vendor, PC's for Everyone. I've both built and purchased desktops. The PC's operation is a very good Linux friendly whitebox vendor. They are also in Microsoft's top 40 manufacturers. I've bought from them and their quality is excellent. So I looked at their laptops.
There are always questions that need answering. Their page is much more complete than someone like Dell, but still I had questions. So, I figured out who made their laptops. That was a fun bit of google searching and image processing. For one laptop I zoomed in and enhanced the photo so that I could read the manufacturers label. It turns out that they sell laptops made by Spartan and Sager. Further google searching leads to some concerns about the Sager. But the problem complaints are old and related to Sager pushing the fringes when cramming advanced capabilities into laptops.
If I go with a Sager, I get the laptop configuration that I want (without OS) for $1750. I might be able to drop that to $1600 if they are willing to remove the CD/RW/DVD and replace it with a plain CD. Sager makes that configuration but PC's doesn't list it. So I won't get an Apple. I get everything I want on a PC for at least $750 less. The display would be 15.1 inch, 1400x1050. That will be usable. The rest is a quite capable 1 GHz Pentium III configuration, using display hardware that is understood by XFree86. There might be some power management issues to track down, but this laptop design does not look too exotic. So Linux should work.
So the next step is determining the cost of caution. Do I go with only Linux? Do I get Windows 2000 for the laptop and set up a dual boot? Do I get a second disk drive so that I can have a Windows drive and a Linux drive? This probably entails talking to PC's and looking at the system. I've found the prices on various locations. Windows 2000 Pro will cost $150. A second disk with bracket will cost about $150. Further complicating this decision is the fact that it will be much harder to get a Windows 2000 laptop in another year. Microsoft is phasing it out.
I do not want to go to Windows XP until EULA's and laws change noticably. Buried in the current EULA for XP (and also in the recent SP3 for Windows 2000) is the agreement that you will permit Microsoft to unilaterally examine your computer, disable or remove software that they dislike, and update other software at their discretion. I will not agree if given a choice. Microsoft PR says that this is just to permit easier security related upgrades, but the license agreement lacks restrictions. Even if it were limited to that, Microsoft's track record with security decisions is rather poor. It is also noteworthy that they disclaim any liability for damages that this might cause.
So the saga will continue.
A couple tropical swirls. The one in the Gulf of Mexico advanced to tropical depression but went inland before getting any further. The one off the Carolinas has a very nice swirl at the moment. This one should make it to tropical storm (and maybe hurricane) before it drifts over colder waters.
For a while I've been lusting after a laptop. This doesn't mean that I will actually go buy one. Half the fun is doing the research, evaluations, and shopping. After all that I may still decide to wait another six months or a year. Then I can do it all over again.
I don't need a laptop. The desktop network has been just fine. But it would be nice to be able to roam around the house instead of being stuck at a desk. Most of my personal work is Unix based, so the laptop choice is more complex. It would be primarily operating as an X-server communicating over SSH encrypted links (over WEP over 802.11b) to the real computer. That continues my practice of preserving the same user view regardless of what computer I am using. It also deals with the security risks of wireless work. So what I need is not CPU power, but X-windows and Unix capability. As a secondary consideration I might take a laptop traveling, but it would primarily be for network access, so the requirements are similar.
I've had corporate laptops for over 10 years. The first one was a 386 laptop, and I've had some kind of laptop ever since. I could use my current employer's laptop, but I practice extremely strict hardware separation of personal from employer use. I think that this is appropriate from an ethical, moral, legal, and practical perspective. That last may need some explanation. I've experienced and witnessed thefts, takeovers, and bankruptcies. There is a very real chance that I might lose the laptop without warning. It would be just suddenly gone and in the hands of someone who might try to use its contents against me. I structure my computer use such that the sudden loss of the corporate laptop would not disturb my personal life.
So I have enough experience to confidently define my ideal laptop:
- Big screen, 15 or 16 inch, at least 1280x1024
- Long battery life
- Unix supported display controller, etc.
- Adequate processor and disk, lots of RAM
- WiFi
- CF access for the digital camera.
Then on to the PC world. Lots of fun looking through web sites. I ended up investigating the ones offered by my favorite vendor, PC's for Everyone. I've both built and purchased desktops. The PC's operation is a very good Linux friendly whitebox vendor. They are also in Microsoft's top 40 manufacturers. I've bought from them and their quality is excellent. So I looked at their laptops.
There are always questions that need answering. Their page is much more complete than someone like Dell, but still I had questions. So, I figured out who made their laptops. That was a fun bit of google searching and image processing. For one laptop I zoomed in and enhanced the photo so that I could read the manufacturers label. It turns out that they sell laptops made by Spartan and Sager. Further google searching leads to some concerns about the Sager. But the problem complaints are old and related to Sager pushing the fringes when cramming advanced capabilities into laptops.
If I go with a Sager, I get the laptop configuration that I want (without OS) for $1750. I might be able to drop that to $1600 if they are willing to remove the CD/RW/DVD and replace it with a plain CD. Sager makes that configuration but PC's doesn't list it. So I won't get an Apple. I get everything I want on a PC for at least $750 less. The display would be 15.1 inch, 1400x1050. That will be usable. The rest is a quite capable 1 GHz Pentium III configuration, using display hardware that is understood by XFree86. There might be some power management issues to track down, but this laptop design does not look too exotic. So Linux should work.
So the next step is determining the cost of caution. Do I go with only Linux? Do I get Windows 2000 for the laptop and set up a dual boot? Do I get a second disk drive so that I can have a Windows drive and a Linux drive? This probably entails talking to PC's and looking at the system. I've found the prices on various locations. Windows 2000 Pro will cost $150. A second disk with bracket will cost about $150. Further complicating this decision is the fact that it will be much harder to get a Windows 2000 laptop in another year. Microsoft is phasing it out.
I do not want to go to Windows XP until EULA's and laws change noticably. Buried in the current EULA for XP (and also in the recent SP3 for Windows 2000) is the agreement that you will permit Microsoft to unilaterally examine your computer, disable or remove software that they dislike, and update other software at their discretion. I will not agree if given a choice. Microsoft PR says that this is just to permit easier security related upgrades, but the license agreement lacks restrictions. Even if it were limited to that, Microsoft's track record with security decisions is rather poor. It is also noteworthy that they disclaim any liability for damages that this might cause.
So the saga will continue.
rjh at 2300
[link]
A couple tropical swirls. The one in the Gulf of Mexico advanced to tropical depression but went inland before getting any further. The one off the Carolinas has a very nice swirl at the moment. This one should make it to tropical storm (and maybe hurricane) before it drifts over colder waters.