While reading the ZDNet article, one thing struck me. They wrote:
The event attracted a wide variety of spectators, including at least one hacker-type who
sported a T-shirt displaying the illegal DeCSS code.
Of course the act of displaying the code on the T-shirt is an act of purely political speech! Nobody really expects people to copy the evil code from the shirt and use it to descramble movies. So here is a very clear example of code devoid of its functional properties, purely used to make a political statement.
For me, the DeCSS gallery by David Touretzky is more than enough to convince me that code is speech, and that code is expressive. When I teach my students, I use both real code and pseudo-code frequently to illustrate important points.
Right. I think a decent solution would be for Mike to match Patrick's $100 and donate the total to the EFF or some other related charity.
And Mike should probably both reword the challenge and, to avoid the appearance of conflicting interests, offer to donate future processing fees (after the callenger admits defeat) in any case.
This argument is spurious. The GPL works, because it grants you more rights than you have under copyright. The PDF copy protection bits, on the other hand, do not stop you from making as many copies as you want - they only stop you from printing the document or from making excepts. Both of these are not necessarily illegal. Most instances of them are covered under fair use - i.e. print a copy onto paper to avoid problems with poor sight (somebody else suggested running pdf2ascii to be able to feed a document into a Braille reader), except parts for literary critique, and so on.
The GPL gives you rights beyond those you usually have. You do not need to "request people to honour the licencse", they are forced by law to honour it (although the law may not always be enforced). The PDF protection bits are at most comments or requests of the author, without any legal backing (except perhaps by the DMCA's stupidity).
Far too much effort is being wasted in superfluous programs that have no real chance of every producing a usable application. For centuries, the common man has supported the scientific elite in their search for 'knowledge', and now I think its time that the debt be paid back. Rather than focusing on theoretical work, itis time for scientists to submit to their natural social role of technology providers.
"The common man", at least in the industrial countries where most research is funded, lives a lot better today than he did 20, 50, or 100 years ago. Technology does not grow out of thin air, it is always backed by basic research. Yes, not all of basic research has yet lead to marketable products. But we do not have a magic looking glass that will tell us just which pieces of knowledge may be useful in the future.
As an example, number theory (the "queen of mathematics") was considered to be pure research, without any useful application until about 15 years ago. Now it forms the basic of most currently used cryptographic applications.
Well, for starters, we wouldn't have free services like AltaVista or Yahoo!, because they rely on banner revenue for the majority of their income.
Well, Google seems to do fine with only very minor and unobtrusive advertisments. Moreover, Altavista started as a research project at Digital, went to become a showcase of Digital, and only went fully commercial after it was spun off.
I can very well imagine such web services to be provided by research institutions or by sponsors. The pure prestige value of being able to say "All of the Web - indexed by SUN computers" is probably worth more than all of the advertising on AltaVista.
Actually, we might want to have a PGPL (Protocol General Public License) for reference implementations, which could be somewhere in between GPL and BSD, and somewhat orthogonal to the LGPL.
Under such a license, the reference implementation could be copied and incorporated into proprietary programs, but (significant) changes would have to be donated to the community.
"Yes, you can have our Kerboros implementation for you proprietary HellOS 666. No, you cannot fork incompatible proprietary versions from it."
Somebody would have to work out the details (obviously, some types of changes need to be legal to allow the integration of code into other programs) and to write it up in Legalese...
No, the GPL is not destroying intellectual property at all. It may be lowering the monetary value of some competing IP, but so so other competitors.
There is a fine distinction in different values of typical "IP" that Microsoft as well as the MPAA try to make us forget: The monetary value of the intellectual property in a pice of GPL'ed software is close to zero (although there are other sources of monetary value to be had, as e.g. Red Hat and SuSE prove). However, the use value of this piece of software for our society is much higher than that of an equivalent piece of proprietary software, because on the one hand, more people can afford to use the software (beer-free) and on the other hand, the software can be used for more purposes (speech-free).
Similarly, the overall economy is not loosing money if people copy music or programs illegally. The money lost to the publisher is offset by the value of the program that the reciepient gains. In fact, the overall economic gain is likely to be positive, since people may accept free copies who would have never paid for the program.
This does not mean that I support illegal copying (I prefer really Free Software for my tools, and I am willing to pay for my games). However, most arguments by publishers about so-called piracy are just garbage.
Did you suggest using gcc on UltraSparc and POWER processors for benchmark comparisons? You've got to be mad. Those are CPUs where the commercial compilers, such as Sun Workshop (SunPro) and xlc (or Visual Age C or whatever IBM calls it these days), frankly kick the crap out of gcc...
I have repeatedly heard this rumour. While there may be true aspects to it, it certainly is not true in general. I have compiled my equational theorem prover E with both the SunPro compiler suite and gcc 2.95.2, with optimization for fastest code used with both compilers, and have not noted a significant difference in running time (and yes, E is CPU bound and eats up cycles like nothing, and it is complex enough that optimization is difficult and helps a lot in general).
These results are for the 32 bit userland on Solaris, for a program that does mostly integer and only a few floating point operations. At least in this case gcc is on par with SUN's "commercial" compilers, with the added bonus that gcc is widely portable (and hence so is code written for it).
You are probably right. There will always be some copies around, and hackers, or even script kiddies, will be able to compile it and to use it to decrypt movies.
But for DeCSS to be really useful to the Linux community or even to the average consumer, it needs to be included in a decent player, and shipped with e.g. Red Hat, Debian, and SuSE. This will not happen with a product thac cannot be legally distributed.
However, the underlying theme is still more important. If code is not protected as speech, you may not be able to talk effectively about cryptography, about bugs in commercial products, about viruses, or about filter programs. You may not be able to publish programs such as Freenet, Napster or Gnuzilla.
In fact, Napster is something that might hurt even your average Windows user...
German, Japan and France use breeding reactors which recycle the spent nuclear fuel. We dont, due to laws passed during the Carter administration.
Germany does not use any breeder technology (the single pilot project has never been productive and will not be continued), and is in fact phasing out nuclear power over the next 30 years, with the first power plants probably going offline within two years..
Many other European countries are also either phasing nuclear power out, or already have done so. This includes the Scandinavian countries and Austria. France is still pushing for nuclear power, though. And due to deregulation of the power market, some of the power production is simply moving to other countries. On the other hand, most power companies are now offerering green power (i.e. electricity produced using regenerative sources only) at a premium.
I have been going to Denmark or Holland during Christmas holidays for years, and it is interesting to see that in the northern flats of Germany, the Netherlands, and Denmark, wind generators have been increasing enourmously.
CD-RW is indeed another story. Most DVD drives can deal with CD-RW, even if they cannot handle CR-R. Apparently, the technology of CD-RW is closer to DVD than the older CD-R's.
I've never seen a hacker compiling a kernel in a plane. Humm... may be Eric Raymond?
I am not Eric Raymond, but
I have done it occasionally - more often, however, on a train, and more often compiling not kernels, but other systems (most likely my equational theorem prover E . And higher speeds also come in handy for LaTeX-ing large documents.
I generally agree that it makes sense to trade speed for longer battery life. However, given that I can work for about 5 hours on my new notebook (P3 600 with SpeedStep, 64 MB Memory under Linux, the current technological compromise is quite ok for me.
Sorry, I don't follow this. Statements that are true about houses may or may not be true about software.
Don't you understand my example or do you doubt its applicability?
In the first case: If you build a house, you typically select a building company a priory, contract the job to them, and (often) pay them as they go. If the company does not deliver, you can try to recover damages, if the company goes bust you are out of luck. This model might work for software.
Now I agree that there are fundamental differences between physical and intellectual property, however, much software has been written following this model. It just has not yet been tried for mass-market software.
Finally, about Steven King: I agree that the model works for him because he is famous. I doubt that it has anything to do with him being rich. And think that e.g. the popularity of Linus or Alan Cox or even RMS show that you can get a reputation for excellence and reliability without a traditional development model.
Nevertheless, it works for houses, so it can work for software.
More into the realm of intellectual property, I think Steven King is at the moment marketing a novel this way - he will only write the next chapter after having received enough money for the last one.
I think these examples show two ways to cope with the difficulty: Watertight contracts and specifications (which would be beneficial for software development anyways) and chosing reliable companies or individuals for contracting development to.
It's true that the number of copies of the software is unlimited. However, the actual process of
developing the software requires time and resources. This is why software costs money --...
You are right. But a reasonable business process would charge most of the money for developing software, not for licensing copies of it.
Note that this would indeed encourage the creation of new software, whle the current licensing model rather encourages people to milk the last bit of money from existing code. It's not the licensing model but the fierce competition that leads to new software at the moment.
Very likely they were the first ones to offer Linux in a box via normal retail channels at all. SLS and Slackware were there before Red Hat, but not available in any store at that time.
Red Hat also is a very good distribution. They have a fairly good packaging system (and brought it to the market first) and they offer distributions not only for Intel, but also for SPARC and Alpha.
Finally, Linus used to use Red hat early on (probably because of the Alpha support), which very likely gave them a boost in popularity.
Over here in Germany, SuSE was the first distribution widely available in shops, and while I have no hard data, it certainly is the dominant player here.
He said that, despite he believes that, despite the fact that there
could conceivably be a legal cause against people for breaking the GPL
in the past, he's not going to pursue that path ever, and urges
everbody to do likewise. How is "You're now guaranteed I'm not going
to ever sue you for this" an attack?
Note that this is not some pettiness on behalf of RMS, but the result
of a fairly keen knowledge of the legal situation. Unfortunatly,
"legal" is not always "logical".
If I take a piece of GPL-ed code and incorporate it into one of my
programs, the result is a derived work of the GPL-ed code. If I
release my code under an incompatible license, I am in violating the
GPL, and are not allowed to distribute the program. This does not
necessarily change if I remove the GPL-ed piece of code - it can
be argued (and many lawyers would gladly argue) that what I have is
not an original work (for which I would have full copyright), but a
derived work of the original program (which violated the GPL) and
hence of the GPL-ed code segment.
And of course it can be argued that the same is true if I then replace
this segment with another one, regardless of the license of the new
piece of code.
The KDE situation is more convoluted, because KDE presumably always
has been released under "GPL with implicit QT link
permission". However, this combined license also is (according to the
FSF's interpretation of the GPL) incompatible with the pure GPL, and
hence in violation of the GPL. Ergo, KDE violates the GPL, and
relinking it with a new library will not make it compliant.
Note that I do not necessarily like this situation. However, unless we
shoot all the lawyers, we need to cope not only with hacker nitpicking
about code, but also with laywer nitpicking about licenses.
Well, as someone else already pointed out, a lot of free software
started either in corporations or, more frequently, in
academia. Important examples of free software that mostly was done in
academia (but by hackers, most of which worked for free or nominal
wages and free computer access) include the many improvements of BSD,
like virtal memory and TCP/IP networking.
If you are talking about the FSF in particular, I do not think that
there exists any other compiler that combines decent optimization with
the extreme portability of gcc. It's no wonder that e.g. NeXT used gcc
to build Objective C.
I suppost that the idea of building applications on top of a language
system (Emacs) also was fairly new and original at that time (and has
since then been copied lots of times, e.g. by AutoCAD).
GNU grep for a long time was the fastest pattern matcher available
anywhere.
And if you compare the GNU tools (from ls to AWK) with their UNIX
counterparts, all of them are more general as well as more powerful
(and ususally faster). They put significant pressure on UNIX vendors
to improve their tool set (and yet, GNU tools are still superior in
most cases).
Often forgotten, but very significant: patch (by Eric Raymond).
And for sheer hack value, Enlightenment has to be mentioned.
...using the Batkey with the keylength of 2048 Batbits. It would take even the Batcomputer (8 Batbits, 3.5 BatMegaHertz, 32 KB ferrit core drum Batmemory - i.e. wild science fiction!) more than 18 years to crack this using brute force. Unfortunately, the Batkey phrase contains 438 occurrences of the letters "Bat".
Well, it was a cool computer. However, it had some failures that made it unattractive.
No disk drive -- at a time where most of the data exchange was still done on floppies.
To many gimmics and to little raw power for the price. Yes, the price was not to bad for the things build into it. But most people did not need e.g. CD quality sound as much as they needed cycles. SUN was a much better buy in this respect
NeXTStep was excellent, but again, not what most people needed at the time. The multi-media age was still some time off, and most researchers were still writing text mode applications (hell, many are still doing it now).
If you check Moody's original article, you will see that he has corrected the numbers. However, he has not dealt with the other systematic errors in his "analysis". I sent him the following letter describing just one problem:
Dear Mr. Moody,
I am writing to you concering your article "Linux Sux Redux",
available at
http://abcnews.go.com/sections/tech/FredMoody/mood y.html. I am fairly
certain that you are getting a large amount of mail about this
article, so I'll try to keep it short. I noted that you corrected the
original misinterpretation of the statistical data from
Bugtraq. However, your "further research" apparently did not include
reading all of Bugtraqs disclaimers.
In your revised article you state "And the NT number is inflated by
BugTraq's inclusion of IE vulnerabilities, since it considers IE part
of the operating system". If you read the Bugtraq disclaimer, you will
note that BugTraq likewise counts bugs in software coming e.g. with
the Red Hat distribution as Linux Bugs. Considering the extremely
large amount of Software included in the typical Linux distribution
(let alone in _all_ distributions covered by BugTraq), the count for
Linux is inflated to a much larger degree than the count for Windows
NT.
I can understand your desire to keep the conclusion of your article intact,
however, the data collected by BugTraq simply does not allow such a
conclusion - as is clearly stated in the disclaimers.
Slashdot Mirror
For me, the DeCSS gallery by David Touretzky is more than enough to convince me that code is speech, and that code is expressive. When I teach my students, I use both real code and pseudo-code frequently to illustrate important points.
And Mike should probably both reword the challenge and, to avoid the appearance of conflicting interests, offer to donate future processing fees (after the callenger admits defeat) in any case.
The GPL gives you rights beyond those you usually have. You do not need to "request people to honour the licencse", they are forced by law to honour it (although the law may not always be enforced). The PDF protection bits are at most comments or requests of the author, without any legal backing (except perhaps by the DMCA's stupidity).
Far too much effort is being wasted in superfluous programs that have no real chance of every producing a usable application. For centuries, the common man has supported the scientific elite in their search for 'knowledge', and now I think its time that the debt be paid back. Rather than focusing on theoretical work, itis time for scientists to submit to their natural social role of technology providers.
"The common man", at least in the industrial countries where most research is funded, lives a lot better today than he did 20, 50, or 100 years ago. Technology does not grow out of thin air, it is always backed by basic research. Yes, not all of basic research has yet lead to marketable products. But we do not have a magic looking glass that will tell us just which pieces of knowledge may be useful in the future.
As an example, number theory (the "queen of mathematics") was considered to be pure research, without any useful application until about 15 years ago. Now it forms the basic of most currently used cryptographic applications.
Well, Google seems to do fine with only very minor and unobtrusive advertisments. Moreover, Altavista started as a research project at Digital, went to become a showcase of Digital, and only went fully commercial after it was spun off.
I can very well imagine such web services to be provided by research institutions or by sponsors. The pure prestige value of being able to say "All of the Web - indexed by SUN computers" is probably worth more than all of the advertising on AltaVista.
Under such a license, the reference implementation could be copied and incorporated into proprietary programs, but (significant) changes would have to be donated to the community.
"Yes, you can have our Kerboros implementation for you proprietary HellOS 666. No, you cannot fork incompatible proprietary versions from it."
Somebody would have to work out the details (obviously, some types of changes need to be legal to allow the integration of code into other programs) and to write it up in Legalese...
There is a fine distinction in different values of typical "IP" that Microsoft as well as the MPAA try to make us forget: The monetary value of the intellectual property in a pice of GPL'ed software is close to zero (although there are other sources of monetary value to be had, as e.g. Red Hat and SuSE prove). However, the use value of this piece of software for our society is much higher than that of an equivalent piece of proprietary software, because on the one hand, more people can afford to use the software (beer-free) and on the other hand, the software can be used for more purposes (speech-free).
Similarly, the overall economy is not loosing money if people copy music or programs illegally. The money lost to the publisher is offset by the value of the program that the reciepient gains. In fact, the overall economic gain is likely to be positive, since people may accept free copies who would have never paid for the program.
This does not mean that I support illegal copying (I prefer really Free Software for my tools, and I am willing to pay for my games). However, most arguments by publishers about so-called piracy are just garbage.
MUDs are a good example, but even better ones are traditional fantasy roleplaying games like AD&D, GURPS, or Shadowrun.
Thanks. Gives me about 2-3 percent - not very much, but it may add up.
I have repeatedly heard this rumour. While there may be true aspects to it, it certainly is not true in general. I have compiled my equational theorem prover E with both the SunPro compiler suite and gcc 2.95.2, with optimization for fastest code used with both compilers, and have not noted a significant difference in running time (and yes, E is CPU bound and eats up cycles like nothing, and it is complex enough that optimization is difficult and helps a lot in general).
These results are for the 32 bit userland on Solaris, for a program that does mostly integer and only a few floating point operations. At least in this case gcc is on par with SUN's "commercial" compilers, with the added bonus that gcc is widely portable (and hence so is code written for it).
You are probably right. There will always be some copies around, and hackers, or even script kiddies, will be able to compile it and to use it to decrypt movies.
But for DeCSS to be really useful to the Linux community or even to the average consumer, it needs to be included in a decent player, and shipped with e.g. Red Hat, Debian, and SuSE. This will not happen with a product thac cannot be legally distributed.
However, the underlying theme is still more important. If code is not protected as speech, you may not be able to talk effectively about cryptography, about bugs in commercial products, about viruses, or about filter programs. You may not be able to publish programs such as Freenet, Napster or Gnuzilla.
In fact, Napster is something that might hurt even your average Windows user...
Germany does not use any breeder technology (the single pilot project has never been productive and will not be continued), and is in fact phasing out nuclear power over the next 30 years, with the first power plants probably going offline within two years..
Many other European countries are also either phasing nuclear power out, or already have done so. This includes the Scandinavian countries and Austria. France is still pushing for nuclear power, though. And due to deregulation of the power market, some of the power production is simply moving to other countries. On the other hand, most power companies are now offerering green power (i.e. electricity produced using regenerative sources only) at a premium.
I have been going to Denmark or Holland during Christmas holidays for years, and it is interesting to see that in the northern flats of Germany, the Netherlands, and Denmark, wind generators have been increasing enourmously.
CD-RW is indeed another story. Most DVD drives can deal with CD-RW, even if they cannot handle CR-R. Apparently, the technology of CD-RW is closer to DVD than the older CD-R's.
I am not Eric Raymond, but I have done it occasionally - more often, however, on a train, and more often compiling not kernels, but other systems (most likely my equational theorem prover E . And higher speeds also come in handy for LaTeX-ing large documents.
I generally agree that it makes sense to trade speed for longer battery life. However, given that I can work for about 5 hours on my new notebook (P3 600 with SpeedStep, 64 MB Memory under Linux, the current technological compromise is quite ok for me.
Don't you understand my example or do you doubt its applicability?
In the first case: If you build a house, you typically select a building company a priory, contract the job to them, and (often) pay them as they go. If the company does not deliver, you can try to recover damages, if the company goes bust you are out of luck. This model might work for software.
Now I agree that there are fundamental differences between physical and intellectual property, however, much software has been written following this model. It just has not yet been tried for mass-market software.
Finally, about Steven King: I agree that the model works for him because he is famous. I doubt that it has anything to do with him being rich. And think that e.g. the popularity of Linus or Alan Cox or even RMS show that you can get a reputation for excellence and reliability without a traditional development model.
More into the realm of intellectual property, I think Steven King is at the moment marketing a novel this way - he will only write the next chapter after having received enough money for the last one.
I think these examples show two ways to cope with the difficulty: Watertight contracts and specifications (which would be beneficial for software development anyways) and chosing reliable companies or individuals for contracting development to.
You are right. But a reasonable business process would charge most of the money for developing software, not for licensing copies of it.
Note that this would indeed encourage the creation of new software, whle the current licensing model rather encourages people to milk the last bit of money from existing code. It's not the licensing model but the fierce competition that leads to new software at the moment.
Red Hat also is a very good distribution. They have a fairly good packaging system (and brought it to the market first) and they offer distributions not only for Intel, but also for SPARC and Alpha.
Finally, Linus used to use Red hat early on (probably because of the Alpha support), which very likely gave them a boost in popularity.
Over here in Germany, SuSE was the first distribution widely available in shops, and while I have no hard data, it certainly is the dominant player here.
Slighly reworded and snarfed!
"Zen is for opening your eyes. Most other religions and occult movements are for putting on cool-looking blindfolds."
Note that this is not some pettiness on behalf of RMS, but the result of a fairly keen knowledge of the legal situation. Unfortunatly, "legal" is not always "logical".
If I take a piece of GPL-ed code and incorporate it into one of my programs, the result is a derived work of the GPL-ed code. If I release my code under an incompatible license, I am in violating the GPL, and are not allowed to distribute the program. This does not necessarily change if I remove the GPL-ed piece of code - it can be argued (and many lawyers would gladly argue) that what I have is not an original work (for which I would have full copyright), but a derived work of the original program (which violated the GPL) and hence of the GPL-ed code segment.
And of course it can be argued that the same is true if I then replace this segment with another one, regardless of the license of the new piece of code.
The KDE situation is more convoluted, because KDE presumably always has been released under "GPL with implicit QT link permission". However, this combined license also is (according to the FSF's interpretation of the GPL) incompatible with the pure GPL, and hence in violation of the GPL. Ergo, KDE violates the GPL, and relinking it with a new library will not make it compliant.
Note that I do not necessarily like this situation. However, unless we shoot all the lawyers, we need to cope not only with hacker nitpicking about code, but also with laywer nitpicking about licenses.
Ups...sorry for the mix-up
If you are talking about the FSF in particular, I do not think that there exists any other compiler that combines decent optimization with the extreme portability of gcc. It's no wonder that e.g. NeXT used gcc to build Objective C.
I suppost that the idea of building applications on top of a language system (Emacs) also was fairly new and original at that time (and has since then been copied lots of times, e.g. by AutoCAD).
GNU grep for a long time was the fastest pattern matcher available anywhere.
And if you compare the GNU tools (from ls to AWK) with their UNIX counterparts, all of them are more general as well as more powerful (and ususally faster). They put significant pressure on UNIX vendors to improve their tool set (and yet, GNU tools are still superior in most cases).
Often forgotten, but very significant: patch (by Eric Raymond).
And for sheer hack value, Enlightenment has to be mentioned.
...using the Batkey with the keylength of 2048 Batbits. It would take even the Batcomputer (8 Batbits, 3.5 BatMegaHertz, 32 KB ferrit core drum Batmemory - i.e. wild science fiction!) more than 18 years to crack this using brute force. Unfortunately, the Batkey phrase contains 438 occurrences of the letters "Bat".
- No disk drive -- at a time where most of the data exchange was still done on floppies.
- To many gimmics and to little raw power for the price. Yes, the price was not to bad for the things build into it. But most people did not need e.g. CD quality sound as much as they needed cycles. SUN was a much better buy in this respect
NeXTStep was excellent, but again, not what most people needed at the time. The multi-media age was still some time off, and most researchers were still writing text mode applications (hell, many are still doing it now).Dear Mr. Moody,
I am writing to you concering your article "Linux Sux Redux", available at http://abcnews.go.com/sections/tech/FredMoody/mood y.html. I am fairly
certain that you are getting a large amount of mail about this
article, so I'll try to keep it short. I noted that you corrected the
original misinterpretation of the statistical data from
Bugtraq. However, your "further research" apparently did not include
reading all of Bugtraqs disclaimers.
In your revised article you state "And the NT number is inflated by BugTraq's inclusion of IE vulnerabilities, since it considers IE part of the operating system". If you read the Bugtraq disclaimer, you will note that BugTraq likewise counts bugs in software coming e.g. with the Red Hat distribution as Linux Bugs. Considering the extremely large amount of Software included in the typical Linux distribution (let alone in _all_ distributions covered by BugTraq), the count for Linux is inflated to a much larger degree than the count for Windows NT.
I can understand your desire to keep the conclusion of your article intact, however, the data collected by BugTraq simply does not allow such a conclusion - as is clearly stated in the disclaimers.
Yours faithfully,
Stephan Schulz