The advantage this concept has is one of sampling scales. We know very little about Mars in terms of its global minerology, etc. I'd say "geography" and "geology" except some pendant would insist that the correct term is "areology" or "gnu/areology" or something.
In any case, landing in one or two spots tells us about those spots, but we'd really like to know about over overall structure of the planet. On Earth, for example, we have big structures like the Canadian Shield. Landing on one spot and then moving around will tell you that, yep, you got granite over here, too. What we'd like to do is sample on a coarse scale, so we could see that a few 100 km away we've got completely different geology.
That's where the gashopper comes in. It's an extremely clever concept. The Martian atmosphere is mostly CO2, and the cycle of boost, glide, land and recharge could go on for a long time. It's a great way to explore a new planet on a scale that's never been done before.
The difference is that running as a non-admin on Windows is a huge pain, as many programs don't play nicely with non-admin accounts. Windows has a huge legacy of "one user per machine" thinking in its applications development history.
That means that many apps will not run well under non-admin accounts on Windows. Try it sometime and see. Talk to any tech-support person and ask what fraction of calls they get due to people trying to run under non-admin accounts (there's been a spate of this lately as folks upgrade to SP 2 and decide to get a bit more serious about security.)
In comparison, I've run Unix of one kind or another since the 80's and have never had to be root to do anything other than install software or do configuration stuff.
Software patents are even more important than patents in other fields, due to the ease with which software techniques can be duplicated.
This is precisely why software should not be patented. Patents give inventors a monopoly in exchange for full disclosure. What are we, the public, getting in exchange for the monopoly we grant to the inventor in the case of an invention that cannot be used without disclosing it? Nothing. Ergo, no patent should be granted.
These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Program, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works. But when you distribute the same sections as part of a whole which is a work based on the Program, the distribution of the whole must be on the terms of this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it.
Thus, it is not the intent of this section to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Program.
In addition, mere aggregation of another work not based on the Program with the Program (or with a work based on the Program) on a volume of a storage or distribution medium does not bring the other work under the scope of this License.
This seems to me to be entirely clear, or at least as clear as the subject-matter permits. If you write a program that is free of GPL'd code, and that can be run without GPL'd code, then it is not under GPL.
A simple test for what constitutes "mere aggregation" is: will my program run without the GPL'd part? If not, it is a derivative work, if so, it is not. If there is an alternative to the GPL'd part that users can substitute, then your program will run without the GPL'd part.
So in the example you cite, if your code can call any other (de)compression tool in place of gzip, you have not created a derivative work.
This is not legal advice, if you were wondering. If it was, I'd be asking money for it!
The design of many existing mass transit systems suggests that moving people efficiently and conveniently is not a goal of mass transit. A friend who used to live in Princeton pointed out this anomaly to me many years ago: Princeton had excellent train service to New York City, and no parking at the train station. His inference was that the people who designed the system really didn't want people to travel. They just hated cars.
I'm not so sure about that--they may just have been really stupid people, a possibility which can never be discounted. However, a quick Google search on "mass transit social agenda" comes up with gems like this:
I hate the fact that my city, San Jose, put in a politically correct light rail system that is not compatible with anything else. It can't share tracks or vehicles with either BART or Amtrak or CalTrain or any other mass transit system. And it doesn't quite reach any airports or terminals of the other nearby mass transit systems. You can take a bicycle on the San Jose Light Rail, but it's not very convenient. If you don't live very near one station and work very near another station, it's very difficult to dream up an idea why you'd want to ride the thing. At the very least, it should have been built on BART standards so that maybe one day it could connect to other systems, but instead it is a totally distinct island of useless rail.
We do have municipal bus routes that include stops at the train station, but the bus route schedules are not synchronized with the train schedules. This is yet another example of how our systems of mass transportation have not reasonably considered the needs of the masses.
It isn't hard to find these sorts of problems, as anyone who has ever ridden a bus to work regularly knows.
Claim 1 is broad for strategic reasons. It pays to file the broadest patent possible, and see what will get by the examiners. Upon examination it is likely that Claim 1 will be narrowed at the insistence of the examiner, but no company wants to patent less than they can, so there is a logical tendency to make the initial claims overly broad.
Conversely, how plausible is it that the Clovis finds of 1936 just happened to constitute the genuinely oldest artefacts of North American settlement?
Consider: let's suppose there were a few thousand early human sites in North America between 50,000 and 13,000 years ago. Random events will by definition have more time to erase evidence of older sites--that's what "older" means: there is more time between then and now.
Now make an extremely sparse sampling of those sites. What are the odds that the oldest site in your sample is anywhere near as old as the oldest site in the underlying data?
The answer, of course, is "extremely low." If you have a solid grasp of sampling statistics you'll see it would be absolutely astonishing if the Clovis Point people were the earliest North American settlers, because the number of sites sampled at the time of their discovery was extremely small. Far from being surprised that older sites have been found, a scientist looking at these data will say, "Well, yeah. We still need to verify the dating, but it's hardly a shock that there are sites far older than the oldest currently known, given the sparseness of the sample."
After all, more than six million patents have been given out since the founding of the US. How many of those do you think are still valid?
There are as of yesterday (November 16th, 2004) 6,820,278 US patents issued. Taking the somewhat fuzzy 20-year rule, 2,337,260 of them are still valid, or 34%!
Furthermore, 50% of all U.S. patents were issued on or after November 12th, 1968--just 36 years ago.
If patenting continues at the current rate (about 500 per day) in a decade or two the majority of issued patents will be still in force.
From the Harvard Law School article: However, most innovations could not practicably be concealed, and competitors were thus free to mimic them.
That is, business process patents are exactly the sort of thing that should not have patents granted on them, as to use them is to reveal them. So there is no need for the public to grant a monopoly to induce the inventor to reveal them.
The most irritating thing about this is to see a NASA article full of sensible units mN, kW and the like, and then report Isp in "seconds", a dimensionally incorrect anachronism from the days when men were men and pounds was the name for both a unit of mass and a unit of force.
The article references one of the traditional justifications for patents: that an inventor is granted a time-limited monopoly in exchange for full disclosure of the invention.
But with regard to software patents, particularly ones like Amazon's one-click patent, there are many inventions that are effectively self-disclosing: if you see that it is done, you know how it is done.
I wonder if it would be possible under U.S. patent law to challenge these patents on this basis? I strongly doubt it, but the very fact that such inventions are patented is a measure of how badly the patent system needs reform.
Ideas are not property, and patents do not grant property rights. They grant monopoly rights in exchange for something else. What is the "something else" in the case of things like the one-click patent? What are we, the public, getting that we would not get otherwise?
Yeah, pretty soon we'll be having people reporting where Atlantis isn't:
"Crypto-archeologists today reported that they found no sign of Atlantis off the coast of North Dakota. This is a striking new development in Atlantis Studies, which have previously shown that Atlantis once covered 98.7% of the Earth's surface area."
In theory, he could have easily based his machine on boolean logic/arithmetic using relays and electricity - but for some reason, chose not to!
The "some reason" is simple: these ideas are not obvious.
They are "obvious" to us because over the past 200 years a lot of very clever people inched their way slowly out on this particular limb and it didn't snap off. It is difficult (but worth-while) to try to immerse yourself in the mind-set of a past era, to ruthlessly censor your own thoughts that use modern concepts. You'll be amazed at how hard it is to think about things without the language and concepts invented for the purpose, and it'll give you a greater appreciation of how difficult it is to invent that language and those concepts from a starting-point of ignorance.
...for believing the commentary in the Slashdot posting of the article. Based on the updated story, it is clear that the original article is not correct in its claim that the images were there previously.
Oh well, being an arrogant jerk makes me feel right at home on slashdot. Sorry about that.
TCP/IP is included on the list of licensed protocols.
"So what?", you may say.
Well...
The Recitals (which is the part of a license agreement that amongst other things lays out what property the licenser owns and is willing to license) declares that the licensee wants to license these protocols "under any applicable intellectual property rights that Microsoft may have"
But...
There is no conceivable scenario in which Microsoft could have any rights to TCP/IP whatsoever.
So why is it included in the agreement?
This would be like my company, whose products use XML parsers, licensing the XML standard to our users. It would be bizzare on the face of it, and such a contract would be in my view very poorly written. Good contracts contain just what they need to contain, and nothing more. Microsoft's lawyers probably know this.
So why exactly did they invest the effort into creating such an extensive list?
This story is not about Microsoft bashing. It is about a very strange license from a very powerful company, which should give us all pause.
I actually once had a brief hallucination that one could create an SGML-based Legal Markup Language (LML) that would allow machine parsing of all laws. It would be fun, because governments would be tied up forever trying to get a totallity of laws to parse, and would therefore do considerably less damage.
The issue isn't evolution vs. intelligent design, the issue is evolution vs. stupid design. If God made Man, God is an idiot.
The eye is a case in point. Retinal detachment... Lousy self-repair... Medicre acuity... Macular degeneration...
Any decent engineer could have fixed all of these problems (many of them are much less of a problem in species other than humans, so we know they can be solved.)
An interesting thought experiment is to imagine what would have happened had we invaded Germany and removed Hitler instead of ceding the Sudetenland to him. People probably would have said we were overstating the threat, etc. Was Saddam as big a threat as Hitler? (Remember, Hitler had no WMD's either,) Maybe not. But if we had removed Hitler when he invaded the Sudetenland, Hitler wouldn't have been as big a threat.
But Germany under Hitler was clearly arming for war. Iraq under Saddam was clearly not.
So perhaps a better question would be: suppose Hitler had not invaded the Sudentenland, or Poland , and England and her allied colonies (not America, who did not enter the war until over two years later even after those invasions took place) had invaded Germany on the pretext that we thought they might someday re-arm and invade Czechoslovakia, Austria, and quite a few other places?
Suppose Chamberlain had lied, and claimed Germany was building new weapons in secret laboratories, and had used that as a pretext for an invasion?
Would you think that that was wrong? Perhaps even evil? I would.
We need someone to exploit one of the many Diebold machine vulnerablities and use it to report every single vote as being one for Nader. They'd only have to hack a few machines to make the problem glaringly obvious. Bonus points for doing it in a "dead-heat" state where the effect on the final election outcome will be impossible to determine.
The point isn't to throw the election, but to show the world unequivocally that we aren't talking about theoretical possibilities, but a serious practical threat to American democracy.
The outcome would be short-term chaos, as the whole U.S. electoral process would be thrown into disrepute, but the long-term result might be to get all major parties to insist on voter-verified, re-countable paper trails, as were used successfully in the recent referendum in Venezuala.
On the other hand, the long-term outcome might be to round up and shoot everyone with the skills to exploit such e-voting vulnerabilities.
It's a bug. Usability issues are bugs, and any behaviour in a program that makes it easy for the user to do the wrong thing is a bug.
Furthermore, it's a grossly obvious bug. Any "default choice" in electoral software is a bug. Default choices should only exist when there is an obvious default. The whole point of a default choice is user convenience--why make them select when we know what they're going to want to do most of the time?
But the goal-in-context of the main e-voting use-case is TO LET THE USER SELECT A CANDIDATE. The existence of a default choice, which allows the user to exit the use-case WITHOUT MAKING A CANDIDATE SELECTION is a serious design error.
Sorry to shout, but this is such a clear and obvious case of how not to do usability engineering it ought to be used as a teaching example in every usability engineering course in the world.
Just as every physics undergrad comes out of school knowing at least "F=ma and you can't push on a rope" every computer science/engineering grad should come out knowing "Never write your own string class and make sure e-voting software forces the user to positively and explicitly select a candidate."
Slashdot Mirror
The advantage this concept has is one of sampling scales. We know very little about Mars in terms of its global minerology, etc. I'd say "geography" and "geology" except some pendant would insist that the correct term is "areology" or "gnu/areology" or something.
In any case, landing in one or two spots tells us about those spots, but we'd really like to know about over overall structure of the planet. On Earth, for example, we have big structures like the Canadian Shield. Landing on one spot and then moving around will tell you that, yep, you got granite over here, too. What we'd like to do is sample on a coarse scale, so we could see that a few 100 km away we've got completely different geology.
That's where the gashopper comes in. It's an extremely clever concept. The Martian atmosphere is mostly CO2, and the cycle of boost, glide, land and recharge could go on for a long time. It's a great way to explore a new planet on a scale that's never been done before.
--Tom
The difference is that running as a non-admin on Windows is a huge pain, as many programs don't play nicely with non-admin accounts. Windows has a huge legacy of "one user per machine" thinking in its applications development history.
That means that many apps will not run well under non-admin accounts on Windows. Try it sometime and see. Talk to any tech-support person and ask what fraction of calls they get due to people trying to run under non-admin accounts (there's been a spate of this lately as folks upgrade to SP 2 and decide to get a bit more serious about security.)
In comparison, I've run Unix of one kind or another since the 80's and have never had to be root to do anything other than install software or do configuration stuff.
--Tom
Software patents are even more important than patents in other fields, due to the ease with which software techniques can be duplicated.
This is precisely why software should not be patented. Patents give inventors a monopoly in exchange for full disclosure. What are we, the public, getting in exchange for the monopoly we grant to the inventor in the case of an invention that cannot be used without disclosing it? Nothing. Ergo, no patent should be granted.
--Tom
This seems to me to be entirely clear, or at least as clear as the subject-matter permits. If you write a program that is free of GPL'd code, and that can be run without GPL'd code, then it is not under GPL.
A simple test for what constitutes "mere aggregation" is: will my program run without the GPL'd part? If not, it is a derivative work, if so, it is not. If there is an alternative to the GPL'd part that users can substitute, then your program will run without the GPL'd part.
So in the example you cite, if your code can call any other (de)compression tool in place of gzip, you have not created a derivative work.
This is not legal advice, if you were wondering. If it was, I'd be asking money for it!
--Tom
I'm not so sure about that--they may just have been really stupid people, a possibility which can never be discounted. However, a quick Google search on "mass transit social agenda" comes up with gems like this:
It isn't hard to find these sorts of problems, as anyone who has ever ridden a bus to work regularly knows.
--Tom
Claim 1 is broad for strategic reasons. It pays to file the broadest patent possible, and see what will get by the examiners. Upon examination it is likely that Claim 1 will be narrowed at the insistence of the examiner, but no company wants to patent less than they can, so there is a logical tendency to make the initial claims overly broad.
--Tom
Conversely, how plausible is it that the Clovis finds of 1936 just happened to constitute the genuinely oldest artefacts of North American settlement?
Consider: let's suppose there were a few thousand early human sites in North America between 50,000 and 13,000 years ago. Random events will by definition have more time to erase evidence of older sites--that's what "older" means: there is more time between then and now.
Now make an extremely sparse sampling of those sites. What are the odds that the oldest site in your sample is anywhere near as old as the oldest site in the underlying data?
The answer, of course, is "extremely low." If you have a solid grasp of sampling statistics you'll see it would be absolutely astonishing if the Clovis Point people were the earliest North American settlers, because the number of sites sampled at the time of their discovery was extremely small. Far from being surprised that older sites have been found, a scientist looking at these data will say, "Well, yeah. We still need to verify the dating, but it's hardly a shock that there are sites far older than the oldest currently known, given the sparseness of the sample."
--Tom
Penguins! After all, where else in nature to we find upright bipedalism and a great affinity for water?
--Tom
...then they laugh at you, then they fight you. Then you win. -- Gandhi
I'd say we're well on our way to stage 3.
--Tom
After all, more than six million patents have been given out since the founding of the US. How many of those do you think are still valid?
There are as of yesterday (November 16th, 2004) 6,820,278 US patents issued. Taking the somewhat fuzzy 20-year rule, 2,337,260 of them are still valid, or 34%!
Furthermore, 50% of all U.S. patents were issued on or after November 12th, 1968--just 36 years ago.
If patenting continues at the current rate (about 500 per day) in a decade or two the majority of issued patents will be still in force.
--Tom
From the Harvard Law School article: However, most innovations could not practicably be concealed, and competitors were thus free to mimic them.
That is, business process patents are exactly the sort of thing that should not have patents granted on them, as to use them is to reveal them. So there is no need for the public to grant a monopoly to induce the inventor to reveal them.
--Tom
The most irritating thing about this is to see a NASA article full of sensible units mN, kW and the like, and then report Isp in "seconds", a dimensionally incorrect anachronism from the days when men were men and pounds was the name for both a unit of mass and a unit of force.
Could we please start reporting Isp in N*s/kg?
--Tom
The article references one of the traditional justifications for patents: that an inventor is granted a time-limited monopoly in exchange for full disclosure of the invention.
But with regard to software patents, particularly ones like Amazon's one-click patent, there are many inventions that are effectively self-disclosing: if you see that it is done, you know how it is done.
I wonder if it would be possible under U.S. patent law to challenge these patents on this basis? I strongly doubt it, but the very fact that such inventions are patented is a measure of how badly the patent system needs reform.
Ideas are not property, and patents do not grant property rights. They grant monopoly rights in exchange for something else. What is the "something else" in the case of things like the one-click patent? What are we, the public, getting that we would not get otherwise?
--Tom
Yeah, pretty soon we'll be having people reporting where Atlantis isn't:
"Crypto-archeologists today reported that they found no sign of Atlantis off the coast of North Dakota. This is a striking new development in Atlantis Studies, which have previously shown that Atlantis once covered 98.7% of the Earth's surface area."
--Tom
In theory, he could have easily based his machine on boolean logic/arithmetic using relays and electricity - but for some reason, chose not to!
The "some reason" is simple: these ideas are not obvious.
They are "obvious" to us because over the past 200 years a lot of very clever people inched their way slowly out on this particular limb and it didn't snap off. It is difficult (but worth-while) to try to immerse yourself in the mind-set of a past era, to ruthlessly censor your own thoughts that use modern concepts. You'll be amazed at how hard it is to think about things without the language and concepts invented for the purpose, and it'll give you a greater appreciation of how difficult it is to invent that language and those concepts from a starting-point of ignorance.
--Tom
People settle their disputes by killing each other (or, more accurately, sending 18 year olds as proxies to kill each other)
This is not quite specific enough: we send 18 year old boys as proxies to kill each other.
Peace doesn't come from treaties. It comes from the realization that war itself is almost never worth fighting.
Again, more specifically, peace comes from valuing the lives of boys more than the objectives to be pursued via war.
--Tom
Oh well, being an arrogant jerk makes me feel right at home on slashdot. Sorry about that.
--Tom
Right now, America's economy is arguably doing better than any other economy on the planet.
Except it isn't able to pay your own government's bills. America's economy is running on credit, which is not a good way to go at all.
--Tom
What part of "Google used to be able to find them" don't you understand?
--Tom
TCP/IP is included on the list of licensed protocols.
"So what?", you may say.
Well...
The Recitals (which is the part of a license agreement that amongst other things lays out what property the licenser owns and is willing to license) declares that the licensee wants to license these protocols "under any applicable intellectual property rights that Microsoft may have"
But...
There is no conceivable scenario in which Microsoft could have any rights to TCP/IP whatsoever.
So why is it included in the agreement?
This would be like my company, whose products use XML parsers, licensing the XML standard to our users. It would be bizzare on the face of it, and such a contract would be in my view very poorly written. Good contracts contain just what they need to contain, and nothing more. Microsoft's lawyers probably know this.
So why exactly did they invest the effort into creating such an extensive list?
This story is not about Microsoft bashing. It is about a very strange license from a very powerful company, which should give us all pause.
--Tom
I actually once had a brief hallucination that one could create an SGML-based Legal Markup Language (LML) that would allow machine parsing of all laws. It would be fun, because governments would be tied up forever trying to get a totallity of laws to parse, and would therefore do considerably less damage.
--Tom
The issue isn't evolution vs. intelligent design, the issue is evolution vs. stupid design. If God made Man, God is an idiot.
The eye is a case in point. Retinal detachment... Lousy self-repair... Medicre acuity... Macular degeneration...
Any decent engineer could have fixed all of these problems (many of them are much less of a problem in species other than humans, so we know they can be solved.)
--Tom
An interesting thought experiment is to imagine what would have happened had we invaded Germany and removed Hitler instead of ceding the Sudetenland to him. People probably would have said we were overstating the threat, etc. Was Saddam as big a threat as Hitler? (Remember, Hitler had no WMD's either,) Maybe not. But if we had removed Hitler when he invaded the Sudetenland, Hitler wouldn't have been as big a threat.
But Germany under Hitler was clearly arming for war. Iraq under Saddam was clearly not.
So perhaps a better question would be: suppose Hitler had not invaded the Sudentenland, or Poland , and England and her allied colonies (not America, who did not enter the war until over two years later even after those invasions took place) had invaded Germany on the pretext that we thought they might someday re-arm and invade Czechoslovakia, Austria, and quite a few other places?
Suppose Chamberlain had lied, and claimed Germany was building new weapons in secret laboratories, and had used that as a pretext for an invasion?
Would you think that that was wrong? Perhaps even evil? I would.
--Tom
We need someone to exploit one of the many Diebold machine vulnerablities and use it to report every single vote as being one for Nader. They'd only have to hack a few machines to make the problem glaringly obvious. Bonus points for doing it in a "dead-heat" state where the effect on the final election outcome will be impossible to determine.
The point isn't to throw the election, but to show the world unequivocally that we aren't talking about theoretical possibilities, but a serious practical threat to American democracy.
The outcome would be short-term chaos, as the whole U.S. electoral process would be thrown into disrepute, but the long-term result might be to get all major parties to insist on voter-verified, re-countable paper trails, as were used successfully in the recent referendum in Venezuala.
On the other hand, the long-term outcome might be to round up and shoot everyone with the skills to exploit such e-voting vulnerabilities.
--Tom
It's a bug. Usability issues are bugs, and any behaviour in a program that makes it easy for the user to do the wrong thing is a bug.
Furthermore, it's a grossly obvious bug. Any "default choice" in electoral software is a bug. Default choices should only exist when there is an obvious default. The whole point of a default choice is user convenience--why make them select when we know what they're going to want to do most of the time?
But the goal-in-context of the main e-voting use-case is TO LET THE USER SELECT A CANDIDATE. The existence of a default choice, which allows the user to exit the use-case WITHOUT MAKING A CANDIDATE SELECTION is a serious design error.
Sorry to shout, but this is such a clear and obvious case of how not to do usability engineering it ought to be used as a teaching example in every usability engineering course in the world.
Just as every physics undergrad comes out of school knowing at least "F=ma and you can't push on a rope" every computer science/engineering grad should come out knowing "Never write your own string class and make sure e-voting software forces the user to positively and explicitly select a candidate."
--Tom