Oh, come on. Mike Tiemann has done more for free software than Mark Ewing (that's not to say anything against Mark, who's done tons). And he was the person who first proved that free software/open source could work in business, by founding Cygnus.
Cygnus probably has more revenue than the old Red Hat. Do you think that all of the talented people from Cygnus should be driven away, by letting all of the old Red Hat people keep the top jobs?
It's amazingly ignorant to suggest that replacing Ewing with Tiemann will harm open source, or, to be even more ridiculous, that Ewing should talk to a lawyer.
Mike Tiemann also did the gcc port to the Sparc and the Motorola 88K, plus much of the work to make gcc efficient on RISC architectures (the first instruction scheduler and delayed branch support).
He does a lot less gcc/g++ development than he used to, though.
The problem with SmartMoney's map is that it focus on market capitalization, which tends to be based on the fantasies of the investors, as opposed to market share (how much of a given market does a given company control)? It doesn't give any insight into areas where there is a lot of competition vs areas where there is almost none.
Patents had nothing to do with the creation of wavelet compression or of fractal image compression. If anything, it has dramatically held back progress in fractal image compression, by creating a bottleneck.
Wavelet compression is a class of sub-band coding. People were already using sub-band coding for audio and images before any of the wavelet coding patents came out; wavelets are a cleaner and more efficient way to do subband coding but that's what it is. There's been truly innovating work in wavelets, notably by people like Ingrid Daubechies, leading to fast wavelet transforms, but most of these folks were motivated by getting publications and tenure, not patents.
Fractal image compression has been held back by patents; they've just created a huge bottleneck. The theory behind the patents is and has been fairly mainstream, so if the patent holders were never born it wouldn't have set progress back much.
Patents only make a difference in cases where substantial investment is needed to produce the invention, that is, it's so expensive that no one would bother to try unless they had some means of getting substantial returns. Nothing in math is ever like that: your expense is pencil, paper, and maybe coffee. Similarly for anything one person can code up on a PC.
New drugs cost millions to bring to market. Semiconductor fabs cost even more (around a billion dollars, no not a typo). Patents in those areas promote progress, by helping people recover their investments.
The Secretary of Labor has overruled OSHA. The ruling is cancelled. See this CNN story.
Re:The importance (or lack thereof) of uptime
on
Linux Kernel 2.2.14
·
· Score: 4
For machines behind firewalls that are performing some task without any problems, it's best in many cases to just leave 'em alone and let them rack up the uptime.
On the other hand, for a visible machine with a static IP address, hosting web pages or other advertised services, you have to keep ahead of the script kiddies. But not all machines are in that category, far from it.
Sun's vision of Java is a Microsoft vision. That is, Sun wants to be Microsoft, by having the same iron grip on the Java platform that Microsoft has on Windows, and then force everyone to pay money to Sun, while enabling more O/S competition. The result of this fantasy would be that Scott McNealy would become the richest man in the world. HP, of course, will not tolerate this. No company that competes with Sun can tolerate this. For Java to be a standard, Sun will have to give up some power and use its greater experience at Java implementation as a competitive advantage, rather than its iron-fisted control of the platform.
Sun is shaping up to be a major enemy of open source, and a very clever one, by coming up with licensing that looks like open source but is not (the important difference being that all money and all control goes to Sun). This is too bad, because Sun in the past has been a good guy.
If HP takes orders from Bill, then why are they a major funder of the Trillian project (porting Linux and the GNU tools to Merced)?
John Gilmore isn't as well-known as he should be. Of the candidates mentioned, he's most deserving of the award.
Things he's either started or made major contributions to include Cygnus, EFF, the alt groups, GNU tar, GNU gdb, Kerberos, BIND, and the Cypherpunks. He's perhaps the most important activist for overturning the US anti-cryptography laws.
Well, Linux is much more of clone of the Unix kernel than the Hurd is, just the same, the Hurd, when complete, will provide a superset of the functionality of Linux (and, if Linus is right, run slower, but we'll see).
So, since you'll be able to run all your Linux applications on the Hurd (when/if it is completed), a Hurd box will look just like a Unix box (run X, Gnome etc). Where it will differ is that there will be applications that can run under the Hurd but not under Linux (or, at least, apps that would need root privs and kernel modules under Linux but could run in user space on the Hurd).
Rick, you're someone I respect, but you richly deserve flaming, maybe more so than anyone I've encountered in a while. When you're wrong, fess up. Don't post bogus defenses.
I'm sorry, but the sloppiness of your history simply can't be justified, especially since you repeatedly just make assertions that are completely bogus. Almost everything you say about gcc is complete nonsense. Your defense that "the facts are somewhat murky" is so weak as to be embarrassing. There is no murkiness at all; tons of people have been involved and know all about it. Your history is not wrong because of differences of opinion. It is wrong because it is wrong. If you got any of this nonsense from ESR, you need to get him educated. ESR hasn't been involved in gcc development in at least the past four years or so, but this hasn't stopped him from declaiming authoritatively about it.
Example: the origin of pgcc. Stallman didn't "ignore repeated requests" for Pentium-specific optimization, the necessary information was trade secret at the time. The gcc maintainers simply did not have the technical information required, and the folks in a position to work on gcc full-time (mainly at Cygnus) had no contracts to do PC work, so both information and resources were lacking. When the Pentium first came out, one had to sign a nondisclosure to get the needed information and the gcc developers couldn't do that. Intel gradually released more information in dribs and drabs, but at the time, instruction scheduling on the Pentium was very tricky and not well understood outside of Intel.
Finally, some Intel engineers did a hack of gcc version 2.4.0 to do Pentium optimization. Unfortunately, they made no attempt to honor the structure of the compiler, e.g. the distinction between the front end (machine-independent) and the back end (machine-dependent). This meant that it wasn't possible to integrate their work. pgcc is based on that work. Thanks to Marc Lehmann and others, the distance between pgcc and egcs/gcc has been steadily reduced; pgcc is currently maintained as a patch against first egcs and then gcc, and the size of the patch has steadily been reduced. Cygnus never had anything to do with pgcc; the Cygnus developers I know considered the original pgcc to be misdesigned and buggy, though it has gotten much better since then.
egcs started as a branch off of the gcc2 development tree, in the long period between 2.7.2 and 2.8.0. I was involved in the discussions that led to the project. When we started egcs, we talked to the pgcc people because we were seeking to decrease the number of gcc forks in the world; in addition to pgcc there were the Cygnus customer releases, which were ahead of the FSF release at the time. I won't go into all the breakdowns that had stalled gcc development at the time, but it had become a mess. We were definitely motivated by ESR's CatB paper. However, we did, in the process of egcs development, demonstrate that one of ESR's contentions is false: copyright assignments are not a barrier to the bizaar model. egcs/gcc is closer to the bizaar model than the Linux kernel, as far more developers have the right to check in code.
With the assistance of Intel, Cygnus has produced a new ia32 backend, which should be out in gcc 2.96; this will finally make pgcc obsolete and hopefully complete the reunification of gcc.
On other topics:
Your history of Lucid Emacs/XEmacs is, as Per has pointed out, completely wrong -- it was GPLed from day one, and much of the code in XEmacs was written by RMS. Go talk to Jamie Zawinski, father of XEmacs, for some education (you might have heard of him;-). Copyright assignments were one issue that kept the fork from healing, but technical differences between RMS and the XEmacs maintainers were also a factor.
Your history of the Unix forks isn't as bad, but it appears that you have the chronology confused a bit. The splits in the BSD camp predated the AT&T/BSDI/UC Berkeley lawsuit, for example.
heroine claims that In 1998 90% of Cygnus's revenue was based on versions of GNU pro and Code fusion which ran on Windows....
Cygnus is (was) a private company and never reported numbers that would enable you to make this estimate, which it appears you are just making up. Cygnus's big-money customers were generally doing embedded systems development, not Windows development, so Cygnus was in no way competing with MSVC++.
I'm a member of the GCC (formerly egcs) steering committee, but this is just me speaking here.
egcs and FSF gcc re-merged in April 1999. There is no longer a separate egcs project. The former egcs team now controls the FSF gcc, the rifts have been patched up, and some of the key GCC developers are competitors to Cygnus (and possibly to Red Hat).
Cygnus marketing has no say on GCC releases, just as they had no say on egcs releases. Red Hat marketing won't be able to affect the timing or quality of GCC 3.0.
Cygnus donates a machine and network bandwidth, plus developers; however, the steering committee, not Cygnus (or Red Hat) controls the direction of the project.
Both Cygnus and Red Hat have good records in their interaction with the free software community, so even if we didn't have the safeguards in place that we do, there wouldn't really be anything to worry about.
Actually, I'm sure that China has been thinking about the very same question, in reverse. They can't see the source to Windows, and for all they know, the NSA has put back doors into every Chinese-language version of Windows. (Those of you who think that the NSA doesn't do this kind of thing, please read this link).
That is, the Chinese know that they can't trust Windows. But the Chinese can't sneak hacks into Linux either, since they have to provide source code and it has to pass review by Linus and the other kernel hackers.
David Brin is right on this stuff
on
RoboFly
·
· Score: 2
David Brin has been pointing out that privacy is going away, because of things like this (causing people to flame him because they think Brin is advocating the loss of privacy, rather than simply predicting it). What's the use of PGP when the fly in the corner is recording your passphrase?
Brin advocates making the lack of privacy democratic: not only can the authorities spy on the public, but anyone in the public can spy on the authorities. No cop would get away with beating a confession out of a suspect ever again, not with hundreds of cameras following his every move.
I don't much like the surveillance society, but I'm afraid that it's coming, and our only choice is to let everyone be a watcher as well as a watchee, or to have a police state the likes of which have never been seen before
For those who think we can pass laws against roboflys, yeah, right. The things will soon be practically invisible, as Moore's Law means that they'll halve in size every couple of years. Laws will be ignored. There will be "arms races" as techologies for detecting micro-spy devices competes with other technology for foiling such detection.
If Red Hat were to pour this kind of money into any of the established groups, they would in effect be buying the group (by being by far the largest contributor).
So it's just as well that they start a separate group. Their foundation is free to make donations to the FSF, Debian, SPI, LSB, etc.
As I type this, there is an animated GIF ad just above the "News for Nerds. Stuff that matters." slogan on my screen. Will Slashot ban GIFs and sacrifice the ad revenue? Or are we about to be embarrassed, thoroughly, when Burn All GIFs day is a bust?
Unfortunately, the organizers didn't do the ground work, like distributing Java or Javascript code that could provide advertisers with alternate means for doing animated ads, plus conversion scripts to instantly turn an animated ad into an alternative form. Yes, this would have required work. But since that work wasn't done, the open source community is about to be embarrassed as every webmaster who depends on ad revenue ignores the call.
Since the FSF, Linus Torvalds and Alan Cox only distribute source code, and since tons of very smart people are studying that code in detail, their room to be sneaky is limited.
If I were the NSA, I wouldn't attempt to corrupt them. I'd get inside Red Hat and f*ck with the binaries they distribute. Who'd know? You may think that you can recompile from source and see if it matches the binaries, but what if they've messed with the compiler (along the lines of the famous Ken Thompson hack)?
They are demanding that a comment with the title "SmartHost for dummies" be stripped from the archives of a mailing list. Do you think that the 1st Amendment has been repealed?
Now, if someone produced a book with a yellow cover called "SmartHost for Dummies", yes, that would be a trademark violation.
Thank you for your letter of October 28. While we understand that it is your job to defend your trademark, we feel that you are asserting an overly broad right to a common phrase in the English language that was in wide use before IDG ever existed as a corporate entity. We do not choose to use the phrase "XXX for dummies" because IDG made it popular; it was a popular phrase before IDG existed as a corporate entity. For this reason, your task of preventing the dilution of the term is probably hopeless; the term is quite diluted already since so many people use it without thinking of IDG books in any way.
While the courts have limited commercial speech, we are not engaging in for-profit activity here and freedom of speech is an important value to us. As private individuals we do not believe that you, or the Congress, has the power to stop us from using the phrase "for dummies" however we choose. We therefore respectfully decline your request.
We assure you that we will not compete with IDG by publishing a book with "... For Dummies" in the title, and we will not in any way imitate the "look and feel" of your "For Dummies" book series.
I don't know why so many moderators think DonkPunch's posting is insightful. Of course authors have the right to choose the licenses they want; it's not like anyone didn't know that before reading DonkPunch's comment.
But this is irrelevant: the original article made a request: that people use one of the well-established licenses, rather than creating new, incompatible ones. Clearly anyone has the absolute right to make whatever requests that they want to make. A surprising number of people have trouble distinguishing requests from orders.
So rights are simply irrelevant here. When you come right down to it, the author has five main choices:
Nearly-public-domain. Let anybody do whatever they want, as long as the author's name stays on.
Require derivative works to also be free software.
Require new versions of the provided code to be free software, but allow other code to link to the free code without it having to be free.
Do one of the previous two, but in an asymmetric way: others have to make their code free, but the author can take users' contributions and make them proprietary.
Use a non-open-source license of some form.
If the author wants to do one of the first three, his/her purpose is accomplished by using a BSD-like license, the GPL, or the LGPL respectively, so there's no point coming up with a new license. If the user wants to do the fourth, it's better than nothing, but it creates problems with license incompatibilities, not just with the GPL, but with other, similar licenses.
The biggest problem is with putting some oddball license on a library, since it may mean that it's not legal to use that library with some other open-source software. I'll bet that a lot of people who want to do something QPL-like could, instead, do something like use the LGPL and also ask people for copyright assignments for changes.
... since the GNOME people are working with the FSF, and the FSF loathes the term "Open Source". (Now I don't care if you want to say "Open Source", but you should know that it's not a universally accepted term).
But yes, GNOME uses the FSF autoconf stuff, which, if used right, creates code that will run on any Unix-like system.
I want to see more programmers get funding to do cool things where they live, rather than feeling they have to come to the US. But it's not up to U.S. VCs to provide funding for this. The problem is the risk-averse culture of those who have money in Europe, Mexico, and elsewhere. It's very difficult to get small high-tech companies started, though I know of a few that have managed to do it (in Belgium and Germany). Many of those are so afraid of looking odd that they try to pretend to be US companies.
There are alternatives, like getting public funding, but often political factors outweigh technical merit (like making sure that each EU country gets its "fair share").
I just slid over to an alternate universe where Linus Torvalds was hit by a bus in 1990, but everything else from 1990 to 1992 was pretty much the same. Thus in this alternate universe there was a free Unix-compatible kernel (actually several: FreeBSD, NetBSD, and Jolitz BSD).
Many of the BSDers jumped to a proprietary company called BSDI. The alternate RMS, fearing a repeat of the Lisp Machines days when all the free software became proprietary again, jumped into action. He kicked the Hurd hackers off the BSD payroll and managed to hire Bill Jolitz, who was so pissed off at the BSDI founders that he abandoned his old opposition to the GPL. RMS put out version 3 of the GPL, which blessed the BSD advertising clause, when he realized that otherwise the GNU/BSD kernel wouldn't be legal.
Now there was a closed group of folks in the GPL camp, and another closed group of folks in the BSD camp. Releases from both of them were eagerly anticipated, but only the elite could play... until along came a group of more democratically minded folks that started a project called, as in this universe, Debian. In an attempt to heal the rift between the two warring camps, they named their project...
Debian GNU/BSD.
Meanwhile the antitrust settlement had split Microsoft into an OS and an application software division, and forced Microsoft to make the MSHTTP spec free (hardly anyone used the web after Microsoft knocked off Netscape in this universe, since they put so many bells and whistles into the spec that no one could write working web pages).
Soon Megasoft Office BSD was out, thanks to porting help from a Scandanavian company named Troll Tech. Within a year, Microsoft (the OS company) had lost significant market share, but every business was forced to buy Megasoft Word, PowerPoint, and Excel. An idealistic group of folks who found this intolerable founded a project called GNOME.
Most folks thought that the Debian and GNOME efforts would fail, but they actually made unexpectedly rapid progress. Eric Raymond became famous when he wrote a paper that documented this.
The biggest danger to Taiwan is people who think like Mr. Conspire. They want to spit in the faces of the most populous country in the world, and then have the US back them up.
It's not going to happen. The only way that Taiwan will get invaded is if it declares independence, since that is not tolerable to the Chinese army. Americans aren't willing to die by the millions to stop such an invasion.
Furthermore the US has signed agreements saying that there is only one China, and that Taiwan is a part of China. After all, that is still the official position of the Taiwanese government! Most Taiwanese people want to be independent, but this has become an issue of pride and face.
Taiwan can remain secure and de-facto independent if they just shut up and stop rocking the boat.
People will, no doubt, attack this message, claiming that it is unprincipled, cowardly, blah blah blah. But how many of you are willing to die so that Taiwan can call itself what it already pretty much is?
If any Taiwanese are under the illusion that, if the Chinese attack the US will save them, they are completely wrong. But don't worry; the Chinese are not insane. As long at you permit them to save their pride, they will not kill the goose that lays the golden eggs. The problem is that you are directly attacking their pride.
Slashdot Mirror
Oh, come on. Mike Tiemann has done more for free software than Mark Ewing (that's not to say anything against Mark, who's done tons). And he was the person who first proved that free software/open source could work in business, by founding Cygnus.
Cygnus probably has more revenue than the old Red Hat. Do you think that all of the talented people from Cygnus should be driven away, by letting all of the old Red Hat people keep the top jobs?
It's amazingly ignorant to suggest that replacing Ewing with Tiemann will harm open source, or, to be even more ridiculous, that Ewing should talk to a lawyer.
Mike Tiemann also did the gcc port to the Sparc and the Motorola 88K, plus much of the work to make gcc efficient on RISC architectures (the first instruction scheduler and delayed branch support).
He does a lot less gcc/g++ development than he used to, though.
The problem with SmartMoney's map is that it focus on market capitalization, which tends to be based on the fantasies of the investors, as opposed to market share (how much of a given market does a given company control)? It doesn't give any insight into areas where there is a lot of competition vs areas where there is almost none.
Patents had nothing to do with the creation of wavelet compression or of fractal image compression. If anything, it has dramatically held back progress in fractal image compression, by creating a bottleneck.
Wavelet compression is a class of sub-band coding. People were already using sub-band coding for audio and images before any of the wavelet coding patents came out; wavelets are a cleaner and more efficient way to do subband coding but that's what it is. There's been truly innovating work in wavelets, notably by people like Ingrid Daubechies, leading to fast wavelet transforms, but most of these folks were motivated by getting publications and tenure, not patents.
Fractal image compression has been held back by patents; they've just created a huge bottleneck. The theory behind the patents is and has been fairly mainstream, so if the patent holders were never born it wouldn't have set progress back much.
Patents only make a difference in cases where substantial investment is needed to produce the invention, that is, it's so expensive that no one would bother to try unless they had some means of getting substantial returns. Nothing in math is ever like that: your expense is pencil, paper, and maybe coffee. Similarly for anything one person can code up on a PC.
New drugs cost millions to bring to market. Semiconductor fabs cost even more (around a billion dollars, no not a typo). Patents in those areas promote progress, by helping people recover their investments.
The Secretary of Labor has overruled OSHA. The ruling is cancelled. See this CNN story.
On the other hand, for a visible machine with a static IP address, hosting web pages or other advertised services, you have to keep ahead of the script kiddies. But not all machines are in that category, far from it.
Sun's vision of Java is a Microsoft vision. That is, Sun wants to be Microsoft, by having the same iron grip on the Java platform that Microsoft has on Windows, and then force everyone to pay money to Sun, while enabling more O/S competition. The result of this fantasy would be that Scott McNealy would become the richest man in the world. HP, of course, will not tolerate this. No company that competes with Sun can tolerate this. For Java to be a standard, Sun will have to give up some power and use its greater experience at Java implementation as a competitive advantage, rather than its iron-fisted control of the platform.
Sun is shaping up to be a major enemy of open source, and a very clever one, by coming up with licensing that looks like open source but is not (the important difference being that all money and all control goes to Sun). This is too bad, because Sun in the past has been a good guy.
If HP takes orders from Bill, then why are they a major funder of the Trillian project (porting Linux and the GNU tools to Merced)?
John Gilmore isn't as well-known as he should be. Of the candidates mentioned, he's most deserving of the award.
Things he's either started or made major contributions to include Cygnus, EFF, the alt groups, GNU tar, GNU gdb, Kerberos, BIND, and the Cypherpunks. He's perhaps the most important activist for overturning the US anti-cryptography laws.
Check out his biography.
Well, Linux is much more of clone of the Unix kernel than the Hurd is, just the same, the Hurd, when complete, will provide a superset of the functionality of Linux (and, if Linus is right, run slower, but we'll see).
So, since you'll be able to run all your Linux applications on the Hurd (when/if it is completed), a Hurd box will look just like a Unix box (run X, Gnome etc). Where it will differ is that there will be applications that can run under the Hurd but not under Linux (or, at least, apps that would need root privs and kernel modules under Linux but could run in user space on the Hurd).
Rick, you're someone I respect, but you richly deserve flaming, maybe more so than anyone I've encountered in a while. When you're wrong, fess up. Don't post bogus defenses.
I'm sorry, but the sloppiness of your history simply can't be justified, especially since you repeatedly just make assertions that are completely bogus. Almost everything you say about gcc is complete nonsense. Your defense that "the facts are somewhat murky" is so weak as to be embarrassing. There is no murkiness at all; tons of people have been involved and know all about it. Your history is not wrong because of differences of opinion. It is wrong because it is wrong. If you got any of this nonsense from ESR, you need to get him educated. ESR hasn't been involved in gcc development in at least the past four years or so, but this hasn't stopped him from declaiming authoritatively about it.
Example: the origin of pgcc. Stallman didn't "ignore repeated requests" for Pentium-specific optimization, the necessary information was trade secret at the time. The gcc maintainers simply did not have the technical information required, and the folks in a position to work on gcc full-time (mainly at Cygnus) had no contracts to do PC work, so both information and resources were lacking. When the Pentium first came out, one had to sign a nondisclosure to get the needed information and the gcc developers couldn't do that. Intel gradually released more information in dribs and drabs, but at the time, instruction scheduling on the Pentium was very tricky and not well understood outside of Intel.
Finally, some Intel engineers did a hack of gcc version 2.4.0 to do Pentium optimization. Unfortunately, they made no attempt to honor the structure of the compiler, e.g. the distinction between the front end (machine-independent) and the back end (machine-dependent). This meant that it wasn't possible to integrate their work. pgcc is based on that work. Thanks to Marc Lehmann and others, the distance between pgcc and egcs/gcc has been steadily reduced; pgcc is currently maintained as a patch against first egcs and then gcc, and the size of the patch has steadily been reduced. Cygnus never had anything to do with pgcc; the Cygnus developers I know considered the original pgcc to be misdesigned and buggy, though it has gotten much better since then.
egcs started as a branch off of the gcc2 development tree, in the long period between 2.7.2 and 2.8.0. I was involved in the discussions that led to the project. When we started egcs, we talked to the pgcc people because we were seeking to decrease the number of gcc forks in the world; in addition to pgcc there were the Cygnus customer releases, which were ahead of the FSF release at the time. I won't go into all the breakdowns that had stalled gcc development at the time, but it had become a mess. We were definitely motivated by ESR's CatB paper. However, we did, in the process of egcs development, demonstrate that one of ESR's contentions is false: copyright assignments are not a barrier to the bizaar model. egcs/gcc is closer to the bizaar model than the Linux kernel, as far more developers have the right to check in code.
With the assistance of Intel, Cygnus has produced a new ia32 backend, which should be out in gcc 2.96; this will finally make pgcc obsolete and hopefully complete the reunification of gcc.
On other topics:
Your history of Lucid Emacs/XEmacs is, as Per has pointed out, completely wrong -- it was GPLed from day one, and much of the code in XEmacs was written by RMS. Go talk to Jamie Zawinski, father of XEmacs, for some education (you might have heard of him ;-). Copyright assignments were one issue that kept the fork from healing, but technical differences between RMS and the XEmacs maintainers were also a factor.
Your history of the Unix forks isn't as bad, but it appears that you have the chronology confused a bit. The splits in the BSD camp predated the AT&T/BSDI/UC Berkeley lawsuit, for example.
heroine claims that In 1998 90% of Cygnus's revenue was based on versions of GNU pro and Code fusion which ran on Windows ....
Cygnus is (was) a private company and never reported numbers that would enable you to make this estimate, which it appears you are just making up. Cygnus's big-money customers were generally doing embedded systems development, not Windows development, so Cygnus was in no way competing with MSVC++.
The Linux box was sub-divided among all of the thousands of folks who told them to leave the name the same. One transistor per person ...
I'm a member of the GCC (formerly egcs) steering committee, but this is just me speaking here.
egcs and FSF gcc re-merged in April 1999. There is no longer a separate egcs project. The former egcs team now controls the FSF gcc, the rifts have been patched up, and some of the key GCC developers are competitors to Cygnus (and possibly to Red Hat).
Cygnus marketing has no say on GCC releases, just as they had no say on egcs releases. Red Hat marketing won't be able to affect the timing or quality of GCC 3.0.
Cygnus donates a machine and network bandwidth, plus developers; however, the steering committee, not Cygnus (or Red Hat) controls the direction of the project.
Both Cygnus and Red Hat have good records in their interaction with the free software community, so even if we didn't have the safeguards in place that we do, there wouldn't really be anything to worry about.
Actually, I'm sure that China has been thinking about the very same question, in reverse. They can't see the source to Windows, and for all they know, the NSA has put back doors into every Chinese-language version of Windows. (Those of you who think that the NSA doesn't do this kind of thing, please read this link).
That is, the Chinese know that they can't trust Windows. But the Chinese can't sneak hacks into Linux either, since they have to provide source code and it has to pass review by Linus and the other kernel hackers.
David Brin has been pointing out that privacy is going away, because of things like this (causing people to flame him because they think Brin is advocating the loss of privacy, rather than simply predicting it). What's the use of PGP when the fly in the corner is recording your passphrase?
Brin advocates making the lack of privacy democratic: not only can the authorities spy on the public, but anyone in the public can spy on the authorities. No cop would get away with beating a confession out of a suspect ever again, not with hundreds of cameras following his every move.
I don't much like the surveillance society, but I'm afraid that it's coming, and our only choice is to let everyone be a watcher as well as a watchee, or to have a police state the likes of which have never been seen before
For those who think we can pass laws against roboflys, yeah, right. The things will soon be practically invisible, as Moore's Law means that they'll halve in size every couple of years. Laws will be ignored. There will be "arms races" as techologies for detecting micro-spy devices competes with other technology for foiling such detection.
If Red Hat were to pour this kind of money into any of the established groups, they would in effect be buying the group (by being by far the largest contributor).
So it's just as well that they start a separate group. Their foundation is free to make donations to the FSF, Debian, SPI, LSB, etc.
As I type this, there is an animated GIF ad just above the "News for Nerds. Stuff that matters." slogan on my screen. Will Slashot ban GIFs and sacrifice the ad revenue? Or are we about to be embarrassed, thoroughly, when Burn All GIFs day is a bust?
Unfortunately, the organizers didn't do the ground work, like distributing Java or Javascript code that could provide advertisers with alternate means for doing animated ads, plus conversion scripts to instantly turn an animated ad into an alternative form. Yes, this would have required work. But since that work wasn't done, the open source community is about to be embarrassed as every webmaster who depends on ad revenue ignores the call.
Since the FSF, Linus Torvalds and Alan Cox only distribute source code, and since tons of very smart people are studying that code in detail, their room to be sneaky is limited.
If I were the NSA, I wouldn't attempt to corrupt them. I'd get inside Red Hat and f*ck with the binaries they distribute. Who'd know? You may think that you can recompile from source and see if it matches the binaries, but what if they've messed with the compiler (along the lines of the famous Ken Thompson hack)?
They are demanding that a comment with the title "SmartHost for dummies" be stripped from the archives of a mailing list. Do you think that the 1st Amendment has been repealed?
Now, if someone produced a book with a yellow cover called "SmartHost for Dummies", yes, that would be a trademark violation.
I suggest a response along the following lines:
Dear Ms. Drewelow,
Thank you for your letter of October 28. While we understand that it is your job to defend your trademark, we feel that you are asserting an overly broad right to a common phrase in the English language that was in wide use before IDG ever existed as a corporate entity. We do not choose to use the phrase "XXX for dummies" because IDG made it popular; it was a popular phrase before IDG existed as a corporate entity. For this reason, your task of preventing the dilution of the term is probably hopeless; the term is quite diluted already since so many people use it without thinking of IDG books in any way.
While the courts have limited commercial speech, we are not engaging in for-profit activity here and freedom of speech is an important value to us. As private individuals we do not believe that you, or the Congress, has the power to stop us from using the phrase "for dummies" however we choose. We therefore respectfully decline your request.
We assure you that we will not compete with IDG by publishing a book with "... For Dummies" in the title, and we will not in any way imitate the "look and feel" of your "For Dummies" book series.
I don't know why so many moderators think DonkPunch's posting is insightful. Of course authors have the right to choose the licenses they want; it's not like anyone didn't know that before reading DonkPunch's comment.
But this is irrelevant: the original article made a request: that people use one of the well-established licenses, rather than creating new, incompatible ones. Clearly anyone has the absolute right to make whatever requests that they want to make. A surprising number of people have trouble distinguishing requests from orders.
So rights are simply irrelevant here. When you come right down to it, the author has five main choices:
If the author wants to do one of the first three, his/her purpose is accomplished by using a BSD-like license, the GPL, or the LGPL respectively, so there's no point coming up with a new license. If the user wants to do the fourth, it's better than nothing, but it creates problems with license incompatibilities, not just with the GPL, but with other, similar licenses.
The biggest problem is with putting some oddball license on a library, since it may mean that it's not legal to use that library with some other open-source software. I'll bet that a lot of people who want to do something QPL-like could, instead, do something like use the LGPL and also ask people for copyright assignments for changes.
But yes, GNOME uses the FSF autoconf stuff, which, if used right, creates code that will run on any Unix-like system.
You write: U.S. VCs, are you listening?.
I want to see more programmers get funding to do cool things where they live, rather than feeling they have to come to the US. But it's not up to U.S. VCs to provide funding for this. The problem is the risk-averse culture of those who have money in Europe, Mexico, and elsewhere. It's very difficult to get small high-tech companies started, though I know of a few that have managed to do it (in Belgium and Germany). Many of those are so afraid of looking odd that they try to pretend to be US companies.
There are alternatives, like getting public funding, but often political factors outweigh technical merit (like making sure that each EU country gets its "fair share").
I just slid over to an alternate universe where Linus Torvalds was hit by a bus in 1990, but everything else from 1990 to 1992 was pretty much the same. Thus in this alternate universe there was a free Unix-compatible kernel (actually several: FreeBSD, NetBSD, and Jolitz BSD).
Many of the BSDers jumped to a proprietary company called BSDI. The alternate RMS, fearing a repeat of the Lisp Machines days when all the free software became proprietary again, jumped into action. He kicked the Hurd hackers off the BSD payroll and managed to hire Bill Jolitz, who was so pissed off at the BSDI founders that he abandoned his old opposition to the GPL. RMS put out version 3 of the GPL, which blessed the BSD advertising clause, when he realized that otherwise the GNU/BSD kernel wouldn't be legal.
Now there was a closed group of folks in the GPL camp, and another closed group of folks in the BSD camp. Releases from both of them were eagerly anticipated, but only the elite could play ... until along came a group of more democratically minded folks that started a project called, as in this universe, Debian. In an attempt to heal the rift between the two warring camps, they named their project ...
Debian GNU/BSD.
Meanwhile the antitrust settlement had split Microsoft into an OS and an application software division, and forced Microsoft to make the MSHTTP spec free (hardly anyone used the web after Microsoft knocked off Netscape in this universe, since they put so many bells and whistles into the spec that no one could write working web pages).
Soon Megasoft Office BSD was out, thanks to porting help from a Scandanavian company named Troll Tech. Within a year, Microsoft (the OS company) had lost significant market share, but every business was forced to buy Megasoft Word, PowerPoint, and Excel. An idealistic group of folks who found this intolerable founded a project called GNOME.
Most folks thought that the Debian and GNOME efforts would fail, but they actually made unexpectedly rapid progress. Eric Raymond became famous when he wrote a paper that documented this.
The biggest danger to Taiwan is people who think like Mr. Conspire. They want to spit in the faces of the most populous country in the world, and then have the US back them up.
It's not going to happen. The only way that Taiwan will get invaded is if it declares independence, since that is not tolerable to the Chinese army. Americans aren't willing to die by the millions to stop such an invasion.
Furthermore the US has signed agreements saying that there is only one China, and that Taiwan is a part of China. After all, that is still the official position of the Taiwanese government! Most Taiwanese people want to be independent, but this has become an issue of pride and face.
Taiwan can remain secure and de-facto independent if they just shut up and stop rocking the boat.
People will, no doubt, attack this message, claiming that it is unprincipled, cowardly, blah blah blah. But how many of you are willing to die so that Taiwan can call itself what it already pretty much is?
If any Taiwanese are under the illusion that, if the Chinese attack the US will save them, they are completely wrong. But don't worry; the Chinese are not insane. As long at you permit them to save their pride, they will not kill the goose that lays the golden eggs. The problem is that you are directly attacking their pride.