Looking for cars, of course I can't help but look first at "The One Number", how many horsepower (BTW, OT, does the rest of the world measure automobile performance in Watts?)
These days, though, I'm very impressed with cars that have a high ratio of horsepower to gas mileage.
In the metric system, I guess that would reduce to kilograms per cubic second.
Maybe there's a similar ranking for peak torque vs gas mileage; it would be interesting to see a ranking of cars on this basis...
The thing that is special about the NASA computer is that it is a single image system
I did parallel code development on Sun SMP boxes. Starting up jobs, seeing what was going on, killing zombies, debugging was all easier on one system than through different boxes you'd have to ssh over to see.
Even though I was using MPI and getting ready for a distributed memory architecture for the really big runs, the development was easier on the SMP box that showed a single system image.
I haven't used things like OpenMOSIX, and Don Becker, early pioneer of Linux ethernet drivers (not many other folks can claim a complete decade of experience with Linux networking), founded a company called Scyld that sells Linux clusters with single system image.
Sometimes it's convenient to see the whole box as if it were one, even though efficient programming dictates that you become aware of the different costs of data access (network, main memory, cache, disk). Practically speaking, developing and running parallel jobs is a higher user productivity proposition on a single system image.
It wasn't clear to me if these phones were simply hardwired pairs, which would mean if you lost a phone that your security would be compromised.
If each phone saved a cache of public keys from potential correspondents, and the user needed to key-in a private key to authenticate, then it would be more intersting.
Lastly, there should be a stegospeech option where the encrypted channel overlays some uninteresting drivel conversation (you know, the kind of conversation that occupies 90% of cellphone bandwidth anyway...)
If Mono works, then Microsofts own products, those of independent system developers and popular games will all be just as good on Linux, OSX, etc. as they will be on Windows. That should make MS very nervous.
As nice as the potential technology is behind.NET and Mono might be, I tend to be more worried that the interfaces to heavily-used class libraries and components will not be sufficiently specified. That is, not all of the behavior will be replicated if you simply drag `n drop your Visual developed application from the latest Longhorn server over to a cheap Linux box. Does that even happen in the Java world, where cross-platform portability is constantly trumpeted?
Differentiating themselves from commodity Linux platforms and leveraging their strength in Windows seems to be more consistent with Microsoft's business objectives.
I think it's good that Google has a competitor to keep them on their toes, honest, etc.
I'd be really curious to know if there is any implicit shading of news happening by use of different technology or explicit policy.
One way would be to do this comparison:
Use Google's search engine to look up URLs that are critical of Google, favorable to Google, etc. and compare to using MS search engine to lookup URLs that critical of Google, favorable to Google, etc.
Use MSN's search engine to lookup URLs that are critical of MSN, favorable to MSN, etc. and compare to Google using Google's search engine to look for, again, exactly the same topics.
This might also be done with regard to favorite wavelengths on the political spectrum, too, to see if there's any differences in returned results that indicate a different political weight (intentional or incidental, as the case may be.)
Maybe a modified SMTP, with public key exchange, and with some checking to see if some particular host's key was signed by some trustworthy signatures.
Two things come to mind, though:
Is the computational overhead of evaluating the trust metric reasonable?
Will public, anonymous webmail services (eg, hotmail) disappear because spammers using them for outbound messages will cause the the whole domain to be viewed with less trust?
More granularity is needed, so not all users of aol.com get painted with the same brush of a bad trust metric.
If "someuser34l67@aol.com" has a trust metric built up by being a long-standing account with a good record of not being rejected by trustworthy mailhosts, then anonymous email would still be possible.
Outbound rate limitations per account would seem to be a good idea, with more limitations on SMTP connections as the user is "newer and less trusted".
It would be a shame if the technology to defeat spam cost us the ability to send anonymous email.
...to be reminded of the costs of re-integration of countries split during the Cold War.
In case you haven't noticed, North Korea is impoverished.
I'll bet the ratio of per capita incomes between North and South Korea in 2003 is more extreme than the ratio between East and West Germany in 1989.
IIRC, West Germany had a couple of years of financial indigestion trying to re-integrate its eastern block counterpart.
Assuming North Korea doesn't use it's disproportionately developed military to attack the South, it will conceivably want to be peacefully re-integrated and re-built.
That task could easily require more than $80.4 billion.
Given all the hype about bio terrorism and the wrenching effects of this hemorrhagic fever, the public tends to think of Ebola as a foremost danger.
Meanwhile, AIDS, which was a big scare two decades ago, has not become an widespread epidemic in developed nations despite having been around a couple of decades, takes a long time for mortality when properly treated with the latest expensive drugs, and "seems to be something that only gays and drug users get". In the public mind, it's not considered much of a danger.
According to the Salt Lake Tribune, Darl McBride reports that, having receive personal threats, he has employed bodyguards for SCO execs.
Re:Going after HP's customers...
on
SCO News Roundup
·
· Score: 1
here
is an indication that they plan to go after a large Linux user sometime in the next 90 days.
Can SCO put off giving IBM any substantial evidence of infringement for that long?
If SCO waits too long to shake down some specific Linux user, their own hand might get called in the interim.
The countersuits could easily push SCO's market cap into negative territory. At which point any of the fools - err - investors holding worthless SCO stock are going to be asking Darl and the other insiders who sold SCO high earlier for some answers.
Where I live, bank branches are asking for thumbprints from non-account holders wanting to cash checks.
This, despite:
The check was written on that bank.
The person can produce a driver's license to verify that they are the payee.
Yes, it's true that it cuts down costs of fraudulent checks that banks must bear. But it also increases risks to check cashers that their special identifier may be misused. What guarantee does the bank provide that the thumbprint won't be used for the single purpose of preventing fraud on that transaction and that it will be destroyed to prevent any possibility of further misuse?
Heavy-handed tactics like this have really driven people to want to use cash more and more.
But imagine if ATM machines used OCR to record the serial numbers of bills dispensed to people and if banks were required to inventory serial numbers of incoming currency, too.
Credit card and debit card transactions have already reduced the proportion of anonymous financial transactions. The technology exists to reduce financial anonymity a lot further.
What I Want to Know
on
SCO News Roundup
·
· Score: 4, Interesting
...is the status of IBM's filings to compel discovery not just from SCO, but with companies investing in SCO.
This could get particularly sticky if SCO's legal team has a strong financial stake in SCO and the outcome.
Attorney/client privilege is pretty strong, but can it be pried apart if there is evidence of, oh, say fraud?
These big American companies do not like open source,
A few big American companies do not like open source.
A larger number of big American companies don't care much whether it's open source or fried noodles, just whether using it reduces their costs.
Evidence is accumulating that it does.
The rest of the world is ahead of the U.S. in discovering this, but some big American companies, who tend to be cautious and conservative as a rule, are noticing that small and medium sized American companies, and isolated departments within their own operations, have reduced their costs by using free software.
if there's one immutable law of economics, it's this: where there is a demand, there will be a supply.
There's a strong demand for Aging Reversal Pills, but the supply hasn't met the demand:).
Things are more complicated. There's a price curve along the abscissa for both supply and demand.
FOSS sits in a very unusual position in that kind of economy for a couple of reasons.
The price is very very low from people demanding the product and potential suppliers are not strongly motivated by profits.
Strictly in terms of supply and demand, there won't be as many potential suppliers of high-quality FOSS for CAD simply because the profit motive is less.
One might be tempted to dismiss FOSS simply on that basis, but that would be premature.
FOSS can be built incrementally as individual suppliers contribute to the work but don't have to foot the whole bill themselves (cost is in time). Then, the staggering cost of creating a cathedral can be broken up into personally-manageable costs of creating hundreds of bazaar stalls.
Another interesting thing about FOSS development insofar as the supply/demand curves are concerned is that individuals that download an inferior product can become suppliers. The GPL maximizes the rate of recycling of good old code into new products near the zero price limit.
This kind of building a better mousetrap for software is peculiar to written works and to published scientific research. For tangible products like automobiles and television sets, there doesn't seem to be the same possibilty of an aftermarket, where an automobile expert can turn a Ford into a Jaguar and everyone in the world can download a Jaguar to replace their Ford.
Whether these schemata are sufficient for someone besides MS to get a suitable XML document to render on the screen or the printed page in exactly the same fashion that MS does?
The reason I ask is that earlier complaints about Word not being an open documented format were directed to an RTF specification at Microsoft.
But the specification was insufficient for anyone who wanted to know how a Word document would be rendered - for that there were hidden rules in Word's codebase, rules that would change over time, or from platform to platfrom (ask anyone on a Mac).
Knowing how to assemble a virus, will hopefully allow us to defend ourselves against them.
As long as the rate at which the virus reproduces and the level of devastation it causes is not too fast or too irreversible.
Consider the effects of some natural virus and other life forms that have been unleased.
A fungus from the Eastern hemisphere pretty well wiped out the American chestnut tree in short order.
Russian thistle, introduced to North America in the 19th century has likewise become endemic, to the point where tumbleweeds are considered an essential ingredient in any Western film set.
Rabbits in Australia, etc., provide some indication of how rapidly reproducing organisms can spread and how much change they can cause.
Do we trust our knowledge of virus mechanics enough to believe that an inadvertent release of "grey goo" can be undone?
To put it another way:
Even if I'm extremely knowledgeable about cars, have built them from scratch, repaired them, etc., is that sufficient assurance I will be able to stop a speeding car running straight at me in time?
The price elasticity of demand of a product is the responsiveness of the quantity demanded for this product to a change in its price. This is by how much will the quantity demanded rise or fall by the effect of a rise or a fall in the price.
If someone controls the supply exclusively, such as Microsoft (or your phone company, electric company, your credit card company, your gasoline/petrol company) because they don't have competitors, then they will take advantage of their position to maximize their profit.
Microsoft could raise or lower the price of Windows by 30% and many corporate IT people and PC manufacturers that pre-bundle Windows would simply pay it. The quantity sold would not change.
Ultimately, price hikes are limited only by what the customers perceive as the cost of migrating to alternatives. Because of their heavy reliance upon the secrets of Windows and Office, their legacy data in proprietary formats, the time invested learning and training on Microsoft applications, most customers perceive the cost of migrating as substantial.
Given another analogy: if you're a parasite it makes no sense to bleed your host dry abruptly. A better strategy is to bleed your host only as much as you can without causing them to actively try to remove you.
To their credit as businessmen, Microsoft is very close to this kind of a balancing optimum revenue point.
Finally, by tying things like IE and WMP into "Windows" - which ought to be nothing more than a device, memory and process manager - they are helping to make migration away from their platform ever more of a difficulty. Users of Exchange, Outlook, Access, SQLServer, while benefitting from the software, are simultaneously growing in dependence upon Microsoft as a sole supplier.
I like the idea of standard protocols for authentication, but with plenty of flexibility built in.
There should be no reason for Jim's Hardware Shack to have access to my full profile of personal information at all.
It should be sufficient that I can locally create a digital check:
my name or handle (and I should be able to create as many or few as I like),
Jim's Hardware Shack's name (or any of the names they want to use),
my secret pin to sign the check or fund transfer request,
an amount,
a time interval in which the transaction can be performed
the name of a server (aka the bank) (say in Cayman Islands) that will vouch for that transaction.
Then, Jim's Hardware Shack need only submit my digitally-signed transaction request to the named server. The named server is the only information that Jim will need to know.
As long as Jim's Hardware Shack trusts the named server to send them the amount of the transaction and "Jim's Hardware Shack" provides them with some registered server at which to dump the funds (I don't need to know where), Jim shouldn't ever even need to know who I am.
Internet banking can be secure and needn't disclose any more information than is absolutely needed.
Jim's Hardware Shack sure as hell doesn't need my blanket credit card number, my One Single True Name, etc..
...many people seem to have trouble understanding...
I've seen before.
Last year.
When Steve Ballmer was asked about Software Assurance and Enterprise Licensing 6.0 and a number of corporate customers were, to say the least, unhappy with the new terms that Microsoft had come out with for its forthcoming products.
How are you explaining better than Steve?
How can quantifiably differentiate the extra value RHEL 3.0 adds to Linux while simultaneously being a good citizen contributing to free and open source software?
Slashdot Mirror
Exactly.
Looking for cars, of course I can't help but look first at "The One Number", how many horsepower (BTW, OT, does the rest of the world measure automobile performance in Watts?)
These days, though, I'm very impressed with cars that have a high ratio of horsepower to gas mileage.
In the metric system, I guess that would reduce to kilograms per cubic second.
Maybe there's a similar ranking for peak torque vs gas mileage; it would be interesting to see a ranking of cars on this basis...
The thing that is special about the NASA computer is that it is a single image system
I did parallel code development on Sun SMP boxes. Starting up jobs, seeing what was going on, killing zombies, debugging was all easier on one system than through different boxes you'd have to ssh over to see.
Even though I was using MPI and getting ready for a distributed memory architecture for the really big runs, the development was easier on the SMP box that showed a single system image.
I haven't used things like OpenMOSIX, and Don Becker, early pioneer of Linux ethernet drivers (not many other folks can claim a complete decade of experience with Linux networking), founded a company called Scyld that sells Linux clusters with single system image.
Sometimes it's convenient to see the whole box as if it were one, even though efficient programming dictates that you become aware of the different costs of data access (network, main memory, cache, disk). Practically speaking, developing and running parallel jobs is a higher user productivity proposition on a single system image.
SGI doesn't really own NUMA
Of course not.
Anyone that's read Cliver Cussler knows the National Underwater and Marine Agency (NUMA), headed by Admiral Sandecker, is part of the U. S. government.
While Dirk Pitt has done most of the work, NUMA has saved the world from catastrophe on numerous occassions.
It wasn't clear to me if these phones were simply hardwired pairs, which would mean if you lost a phone that your security would be compromised.
If each phone saved a cache of public keys from potential correspondents, and the user needed to key-in a private key to authenticate, then it would be more intersting.
Lastly, there should be a stegospeech option where the encrypted channel overlays some uninteresting drivel conversation (you know, the kind of conversation that occupies 90% of cellphone bandwidth anyway...)
If Mono works, then Microsofts own products, those of independent system developers and popular games will all be just as good on Linux, OSX, etc. as they will be on Windows. That should make MS very nervous.
As nice as the potential technology is behind .NET and Mono might be, I tend to be more worried that the interfaces to heavily-used class libraries and components will not be sufficiently specified. That is, not all of the behavior will be replicated if you simply drag `n drop your Visual developed application from the latest Longhorn server over to a cheap Linux box. Does that even happen in the Java world, where cross-platform portability is constantly trumpeted?
Differentiating themselves from commodity Linux platforms and leveraging their strength in Windows seems to be more consistent with Microsoft's business objectives.
I think it's good that Google has a competitor to keep them on their toes, honest, etc.
I'd be really curious to know if there is any implicit shading of news happening by use of different technology or explicit policy.
One way would be to do this comparison:
- Use Google's search engine to look up URLs that are critical of Google, favorable to Google, etc. and compare to using MS search engine to lookup URLs that critical of Google, favorable to Google, etc.
- Use MSN's search engine to lookup URLs that are critical of MSN, favorable to MSN, etc. and compare to Google using Google's search engine to look for, again, exactly the same topics.
This might also be done with regard to favorite wavelengths on the political spectrum, too, to see if there's any differences in returned results that indicate a different political weight (intentional or incidental, as the case may be.)better off looking for a technical solution
Maybe a modified SMTP, with public key exchange, and with some checking to see if some particular host's key was signed by some trustworthy signatures.
Two things come to mind, though:
More granularity is needed, so not all users of aol.com get painted with the same brush of a bad trust metric.
If "someuser34l67@aol.com" has a trust metric built up by being a long-standing account with a good record of not being rejected by trustworthy mailhosts, then anonymous email would still be possible.
Outbound rate limitations per account would seem to be a good idea, with more limitations on SMTP connections as the user is "newer and less trusted".
It would be a shame if the technology to defeat spam cost us the ability to send anonymous email.
...to be reminded of the costs of re-integration of countries split during the Cold War.
In case you haven't noticed, North Korea is impoverished.
I'll bet the ratio of per capita incomes between North and South Korea in 2003 is more extreme than the ratio between East and West Germany in 1989.
IIRC, West Germany had a couple of years of financial indigestion trying to re-integrate its eastern block counterpart.
Assuming North Korea doesn't use it's disproportionately developed military to attack the South, it will conceivably want to be peacefully re-integrated and re-built.
That task could easily require more than $80.4 billion.
Given all the hype about bio terrorism and the wrenching effects of this hemorrhagic fever, the public tends to think of Ebola as a foremost danger.
Meanwhile, AIDS, which was a big scare two decades ago, has not become an widespread epidemic in developed nations despite having been around a couple of decades, takes a long time for mortality when properly treated with the latest expensive drugs, and "seems to be something that only gays and drug users get". In the public mind, it's not considered much of a danger.
But AIDS is devastating Africa these days.
6-10 Kenyan soldiers die weekly; 80% infected
AIDS orphans outcast
I only wish the price of these things would slide down a little more.
Cost of this 1 teraflop Mellanox machine is less than US$1e6 according to this brochure.
That's considerably less than the US$50e6 that the first teraflop machine cost (Sandia's ASCI Red see this SC1996 flier) 7 years ago.
I don't have a spare million, either, but that kind of 98% price reduction is still fairly impressive.
According to the Salt Lake Tribune, Darl McBride reports that, having receive personal threats, he has employed bodyguards for SCO execs.
here is an indication that they plan to go after a large Linux user sometime in the next 90 days.
Can SCO put off giving IBM any substantial evidence of infringement for that long?
If SCO waits too long to shake down some specific Linux user, their own hand might get called in the interim.
The countersuits could easily push SCO's market cap into negative territory. At which point any of the fools - err - investors holding worthless SCO stock are going to be asking Darl and the other insiders who sold SCO high earlier for some answers.
I had to give a thumbprint.
Where I live, bank branches are asking for thumbprints from non-account holders wanting to cash checks.
This, despite:
- The check was written on that bank.
- The person can produce a driver's license to verify that they are the payee.
Yes, it's true that it cuts down costs of fraudulent checks that banks must bear. But it also increases risks to check cashers that their special identifier may be misused. What guarantee does the bank provide that the thumbprint won't be used for the single purpose of preventing fraud on that transaction and that it will be destroyed to prevent any possibility of further misuse?Heavy-handed tactics like this have really driven people to want to use cash more and more.
The fun side of money tracing is wheresgeorge.com
But imagine if ATM machines used OCR to record the serial numbers of bills dispensed to people and if banks were required to inventory serial numbers of incoming currency, too.
Credit card and debit card transactions have already reduced the proportion of anonymous financial transactions. The technology exists to reduce financial anonymity a lot further.
...is the status of IBM's filings to compel discovery not just from SCO, but with companies investing in SCO.
This could get particularly sticky if SCO's legal team has a strong financial stake in SCO and the outcome.
Attorney/client privilege is pretty strong, but can it be pried apart if there is evidence of, oh, say fraud?
These big American companies do not like open source,
A few big American companies do not like open source.
A larger number of big American companies don't care much whether it's open source or fried noodles, just whether using it reduces their costs.
Evidence is accumulating that it does.
The rest of the world is ahead of the U.S. in discovering this, but some big American companies, who tend to be cautious and conservative as a rule, are noticing that small and medium sized American companies, and isolated departments within their own operations, have reduced their costs by using free software.
I could take an axe to my server room, and reps. from the various vendors would be here onsite in the middle of the night within 45 minutes.
I can verify this is true.
if there's one immutable law of economics, it's this: where there is a demand, there will be a supply.
There's a strong demand for Aging Reversal Pills, but the supply hasn't met the demand:).
Things are more complicated. There's a price curve along the abscissa for both supply and demand.
FOSS sits in a very unusual position in that kind of economy for a couple of reasons.
The price is very very low from people demanding the product and potential suppliers are not strongly motivated by profits.
Strictly in terms of supply and demand, there won't be as many potential suppliers of high-quality FOSS for CAD simply because the profit motive is less.
One might be tempted to dismiss FOSS simply on that basis, but that would be premature.
FOSS can be built incrementally as individual suppliers contribute to the work but don't have to foot the whole bill themselves (cost is in time). Then, the staggering cost of creating a cathedral can be broken up into personally-manageable costs of creating hundreds of bazaar stalls.
Another interesting thing about FOSS development insofar as the supply/demand curves are concerned is that individuals that download an inferior product can become suppliers. The GPL maximizes the rate of recycling of good old code into new products near the zero price limit.
This kind of building a better mousetrap for software is peculiar to written works and to published scientific research. For tangible products like automobiles and television sets, there doesn't seem to be the same possibilty of an aftermarket, where an automobile expert can turn a Ford into a Jaguar and everyone in the world can download a Jaguar to replace their Ford.
Whether these schemata are sufficient for someone besides MS to get a suitable XML document to render on the screen or the printed page in exactly the same fashion that MS does?
The reason I ask is that earlier complaints about Word not being an open documented format were directed to an RTF specification at Microsoft.
But the specification was insufficient for anyone who wanted to know how a Word document would be rendered - for that there were hidden rules in Word's codebase, rules that would change over time, or from platform to platfrom (ask anyone on a Mac).
We don't want you to think, we want you code!!
Reminds me of the line in Beyond Thunderdome, when Mel Gibson is shoveling pigshit, begins to blurt out
and gets wacked with an admonition -I often forget the subtle jabs and they are the most important part of any writing
Unconventional performance, but you did get your subtle jabs in there at the end.
Your warped humor wasn't completely lost.
to get a near-perfect score
Follow the old adage:
Knowing how to assemble a virus, will hopefully allow us to defend ourselves against them.
As long as the rate at which the virus reproduces and the level of devastation it causes is not too fast or too irreversible.
Consider the effects of some natural virus and other life forms that have been unleased.
A fungus from the Eastern hemisphere pretty well wiped out the American chestnut tree in short order.
Russian thistle, introduced to North America in the 19th century has likewise become endemic, to the point where tumbleweeds are considered an essential ingredient in any Western film set.
Rabbits in Australia, etc., provide some indication of how rapidly reproducing organisms can spread and how much change they can cause.
Do we trust our knowledge of virus mechanics enough to believe that an inadvertent release of "grey goo" can be undone?
To put it another way:
How bad does it have to get
It's all a matter of price elasticity.
If someone controls the supply exclusively, such as Microsoft (or your phone company, electric company, your credit card company, your gasoline/petrol company) because they don't have competitors, then they will take advantage of their position to maximize their profit.
Microsoft could raise or lower the price of Windows by 30% and many corporate IT people and PC manufacturers that pre-bundle Windows would simply pay it. The quantity sold would not change.
Ultimately, price hikes are limited only by what the customers perceive as the cost of migrating to alternatives. Because of their heavy reliance upon the secrets of Windows and Office, their legacy data in proprietary formats, the time invested learning and training on Microsoft applications, most customers perceive the cost of migrating as substantial.
Given another analogy: if you're a parasite it makes no sense to bleed your host dry abruptly. A better strategy is to bleed your host only as much as you can without causing them to actively try to remove you.
To their credit as businessmen, Microsoft is very close to this kind of a balancing optimum revenue point.
Finally, by tying things like IE and WMP into "Windows" - which ought to be nothing more than a device, memory and process manager - they are helping to make migration away from their platform ever more of a difficulty. Users of Exchange, Outlook, Access, SQLServer, while benefitting from the software, are simultaneously growing in dependence upon Microsoft as a sole supplier.
I like the idea of standard protocols for authentication, but with plenty of flexibility built in.
There should be no reason for Jim's Hardware Shack to have access to my full profile of personal information at all.
It should be sufficient that I can locally create a digital check:
Then, Jim's Hardware Shack need only submit my digitally-signed transaction request to the named server. The named server is the only information that Jim will need to know.
As long as Jim's Hardware Shack trusts the named server to send them the amount of the transaction and "Jim's Hardware Shack" provides them with some registered server at which to dump the funds (I don't need to know where), Jim shouldn't ever even need to know who I am.
Internet banking can be secure and needn't disclose any more information than is absolutely needed.
Jim's Hardware Shack sure as hell doesn't need my blanket credit card number, my One Single True Name, etc..
The phrase:
I've seen before.Last year.
When Steve Ballmer was asked about Software Assurance and Enterprise Licensing 6.0 and a number of corporate customers were, to say the least, unhappy with the new terms that Microsoft had come out with for its forthcoming products.
How are you explaining better than Steve?
How can quantifiably differentiate the extra value RHEL 3.0 adds to Linux while simultaneously being a good citizen contributing to free and open source software?