"The biggest advantage of Type Ahead Find is that it makes keyboard navigation practical."
Hear, hear! Watch someone using a computer, and see how time he wastes switching from keyboard to mouse. Especially for text searches. Old way: hitting C-f, finding the dialog with the mouse and getting to the text field, moving the dialog so you can see the page, and so on. New way: hit / and start typing. No mousing, no annoying dialog box. Amazing! Years of using "less" has conditioned me to use / for searches, and this works great.
"He in a way is as dangerous as SCO because he is not exactly proping up Linux or IBM!"
I will agree with the people who think he is propping up Linux. He is telling people that they don't have to go out and buy Sun or Microsoft software, because most of their free software is not under any threat. There is no need to get rid of all your Linux servers today, because you won't have to throw out your entire system if SCO should win; the worst case is that you have to get a different more or less equivalent kernel.
I think this is a voice in support in of Linux and free software for the ears of the PHBs.
"If I were sitting on the jury for that case, that's certainly how I would see it."
That is why, under the US court system, you would never be sitting in such a jury. If this case went to a jury (I don't know if it could or not), no programmer or anyone with Linux or Unix experience would be allowed on the jury. You are guaranteed trial by a jury of the clueless.
There was a piece in the Sunday NYT WiR complaining about overzealous use of the word "Orwellian" (also that it refers to what the author described rather than what he espoused, but that is a separate issue). A main point was that every minor restriction in access to information or political euphemism should not be compared to the total control of language and information in 1984.
When I recently reread Farhenheit 451, I caught a detail that I never paid attention to before. The firemen did not start burning books because the government wanted to eliminate them. The people demanded that thought-provoking, controversial and therefore disruptive works be destroyed. For any book, you could find some group who was bothered by it, so all books came under the kerosene.
I see aspects of a sort of reverse-Orwellian society today. Varied viewpoints and honest criticism may be available, but most people don't want to hear them, any more than Bradbury's society wanted their books. Given the choice between the happy myth and reality, people will choose the myth. How many Americans care about the truth of Iraqi WMD's, the Lynch "rescue", or whether Iraq was involved in the 9/11 attacks? Or that fuel cells are not an energy source? Far too few for my comfort.
This is worse than 1984, which envisions domination under an authoritarian government (as I remember; it has been a while). You don't have to beat down the people. Just tell the people the lies they want to hear, and they will do the rest. Whatever contradicts what people already want to believe will be ignored.
This is something fascinating about information access and the internet. The net does not serve to widely disseminate information, except in the most literal geographic sense. Instead, it allows people to form communities with others who already share the same opinions. Memes bounce around in a mostly closed community, building up power and credibility.
I can think of one concrete example. I received a forwarded email in 2000 of stupid statements allegedly made by Al Gore. I replied with an email from 1992 with the same set of quotes, but attributed to Dan Quayle. Did the original sender feel humiliated and send an apologetic retraction to everyone he had forwarded the message to? Of course not. The truth was easily available, but they liked the lie better.
" First of all, file the report. Ask the support person if you can fax in the report because you don't want to inform the hacker that (s)he's been spotted and you are reasonably clear that you can't get a secure channel to their web server."
What am I missing? The problem is with DNS. Use the numerical IP address of the webserver or mail server, and communications should proceed normally.
" Automatic versioning of files is one thing I actually miss from VMS. Certainly not most of the rest of it:)"
I agree completely. As an aside, when you want this kind of versioning, you can get emacs to write out versions as somestuff.~1~, somestuff.~2~,...with the most recent copy named just somestuff. The biggest reason I don't use this is because I can't get ls to sort them the way I would like, especially when it gets to version ~10~.
'And finally; "Both suites of programs contained the same spelling mistakes in the comment lines and the same redundant code. The judge did not accept the argument that this was due to programming style."'
I can see it now... comments in both code misspelled it's for it and loose for lose, used fsck as an expletive, and referred to some mysterious substance known as pr0n.
IBM will then subpoena 600,000 registered/. users to prove that geeks can't spell.
"This seems to me to be a fallacy of composition... Just because she hasn't fully understood the GPL (she is, after all, just a lady who writes articles for a living -- surprisingly throughout this whole SCO issue she has done very little analytical thinking for being one who calls herself an "analyst") doesn't mean that everything else she said is automatically a misrepresentation, or patently false."[snip]
It is not a fallacy to say that a source which provides erroneous information should not be relied upon.
I can see a fallacy if I used the fact of one false statement to conclude that other statements by her are probably false. It is however appropriate to say that I cannot a priori assume that her other statements are either true or false. Normally, you assume what someone says is true, especially if they claim to be an expert; Didio's clear misunderstanding of the GPL means that she no longer deserves to get the "benefit of the doubt." Her other statements must be evaluated on their own merit, just like any random/. post, rather than regarded as informed, expert opinion.
Nothing like an outrageous subject line to attract attention:-)
Didio: "Check the GPL and look at Section 0. It reads that the legal copyright holder of the source code has to explicitly put an assignment and copyright transfer notice into the beginning of the GPL. There is no concept of accidentally giving away the code to the GPL."
That's not what my copy of the GPL says. It says "This License applies to any program or other work which contains a notice placed by the copyright holder saying it may be distributed under the terms of this General Public License."
Further copying and modification are separate. The original author of the program placed the program under the GPL, and it and all distributed derivative works are likewise under the GPL.
If someone modifies and distributes the code, they don't in any way have to reaffirm the GPL notice. They certainly are not required to add additional copyright notices.
Didio's comment about the GPL is a complete misrepresentation, and she reveals herself here as an SCO pawn. I have no reason to believe anything else she said.
It shouldn't be that hard to get the OS certified as official Unix. The difference between a BSD and a legal Unix can't be that much. Many of the important parts may be hidden from the GUI, but that shouldn't matter.
If Apple does this, they can sell their OS as an actual Unix. This would seem to be great as an advertising angle. It is one thing to know that an Apple machine would make a good server because it is Unix-like underneath, but it would have a great impact if it could say Unix right on the box. People who need a new Unix server may start thinking of Apple first.
They have already switched over to a Unix-y system, so why not make the most of it?
"This would not be optimum for SCO, because they could not then demand licensing from all Linux users."
They cannot do this without throwing away their veil of secrecy.
I have source to at least 7 different kernel versions sitting around, mostly old RH cds. They would need to tell me which ones they want licensing fees for. And if their code is not compiled into the base kernel for all possible kernel configurations, then they would need to tell me what combinations of options cause their code to be included in the kernel. After all, if it is only in say jfs and smp code, I won't have it in my kernel, I won't be using it, and I won't owe them any licensing fee.
This is one difference open source makes. If the code were in a commercial product, they could announce which versions infringe. They would need to be much more specific here.
If they do indeed give out the specific information of what kernel versions and what modules infringe, it would be trivial to ferret out the offending code and be done with them.
Re:So let me get this straight...
on
SCO SCO SCO!
·
· Score: 1
"Basically, SCO needs to prove that the code does belong to SCO, and that it predates the "copied" code in Linux. Both require the context of the rest of SCO's source tree."
This statement is true, but I would add that some supposedly infringing code could be proven to not come from SCO.
If SCO says blahblah.c is from their IP, and blahblah.c can be shown to have been lifted from BSD or to have grown from 50 different patches submitted to lkml by 50 different people, then their claim is demonstrably false.
If blahblah.c was mailed whole to Linus from someone at IBM in the right time period, only then would a comparison with SCO's source be necessary.
If they showed the part of the kernel code that they thought infringed, and it appeared even just a little bit possible that they were in the right, I think a big chunk of geek public opinion would swing their way. Why? Because the karma of the whole open source/free software community is at stake here, and people would not take kindly to IBM fscking up that karma.
"Because there is a large number of potential users who will use something else if they can't work out how the thing works in less than five minutes."
I see your point, but I think this is too simplistic and not completely accurate.
I have seen people spend hours of confusion trying to figure out how to do something in MSWord, or some power point thingy, or Eudora. They didn't switch to different software.
There is a clear distinction between usability and marketability. Obviously commercial software has to be made marketable. My point is that so-called usability studies are often really marketability studies. This confusion prevents people from realizing that they can work to make the product both more usable and more marketable. The mistake is to unnecessarily sacrifice usability for marketability.
An obvious outgrowth of this thinking is programs with a sort of beginner mode and an expert mode. This is not uncommon under the *nix philosophy, because the gui and the guts are well separated and configurability is a goal. Any user can use the GUI, but there are command-line modes or ways of turning menus off.
"Maybe I'm crazy but that's one of the few things I like about windows: walk up to every windows machine and know exactly where to go to get what..."
Why should I care whether you can walk up to my machine and use it easily? (That's why I use xlock in the first place.)
Why do people concerned about usability spend so much time worrying about what people can do in the first 5 minutes after they sit down at a 'puter and so little time worrying about what people will be able to accomplish over the next five years?
Once when I was in a grocery store with a friend who spoke Russian, he overheard a conversation two recent arrivals were having. They were annoyed and complaining about the ice cream, because there were too many flavors to choose from.
It seems like the line between enough choice and too much choice isn't as concrete and obvious as some would think.
Re:Why not just....
on
fvwm Turns Ten
·
· Score: 2, Interesting
" marginally useful? what else could you possibly
need. you can move, resize, iconify, and uniconfy
windows."
I need multiple virtual screens. Other than that, I probably could get most of the functionality I want out of twm with a well-written.twmrc. Looking at the man page, there are a lot of useful functions that aren't available unless you customize. For example, I use the fvwm equivalent of TwmWindows frequently; I didn't realize TwmWindows existed because you can't get to it in the default configuration.
Re:Reparenting window managers are for wimps
on
fvwm Turns Ten
·
· Score: 1
All we had to use for punchcards were rocks.
And we had to chew the holes into them with our teeth.
Re:speaking of old window managers
on
fvwm Turns Ten
·
· Score: 4, Insightful
"I was surprised to find twm when I installed X11 on OS X."
I have, more than once, been incredibly relieved to find twm installed as a part of X on machines (not OS 10). Because when the install fails without getting all 90 billion parts of gnome or kde installed correctly, or using an old machine that can't handle the latest and greatest, I can use twm as a marginally useful window manager to start getting things done.
And when this happens, the one of the first things I do is download and install fvwm. Woohoo!
Re:Questions I'd like the experts to answer
on
Today's SCO News
·
· Score: 1
" SCO claims that they will soon reveal the alledged copyright infractions in Linux to a panel of experts. I haven't seen the names of these experts, so I'm a little worried about how expert they may be."
Consider that any expert that sees the "secret evidence" may then be unable to do any future work on the Linux kernel. This means the expert almost certainly cannot be someone familiar with the Linux kernel.
I think that this is SCO's secret weapon (in their minds). If the public knew what the supposed violations were, Linus, Eric Raymond, Bruce Perens, the FSF, and hundreds of others would be submitting amicus curiae briefs blowing SCO's case out of the water. They absolutely need to keep any reliable expert away from their "evidence".
If they accomplish this, then their lawyers and captive "experts" will try to dizzy the judge with techno-BS and win through Proof by Confusion. This is the only reason I can conceive of for keeping secret the location of the alleged violations.
"Where is the logic in keeping the outside experts under NDA about what code is believed by SCO to have been copied into Linux?"
Here's an idle speculation: They know their claims would be slaughtered if exposed to public scrutiny. Dozens of programmers could prove that the material wasn't taken from UNIX. But maybe they have an "expert" who can make a plausible case to a judge. To analyze the code for the court case, any Linux developer would have to sign an NDA essentially prohibiting her from ever working on the kernel again. And SCO thinks no developer with enough knowledge to disprove their claims would be willing to make the sacrifice.
It's a funny second-order trade secret argument. All of the code is public, but somehow they want the right to keep secret the knowledge of which parts of it used to be secret.
How much electrical power production do we need to switch to all electrical (using H2 as an energy storage mechanism)?
What I found on the web says that a car moving at highway speeds uses about 15 kW of power. The standard estimate for domestic power use is 1 kW averaged throughout the day.
Back of the envelope, let's say 10 million Californicators spend an hour a day in their cars. Averaged over 24 hours, this is over 6 GW. Entire daytime power usage in CA is about 35 GW (depending on season). And this doesn't account for SUVs using more power or commercial trucking.
I would be interested in seeing a real estimate, but it looks like this would require a substantial increase in power production facilities.
And this leads to a sticky question. If we can provide electricity via renewables to generate hydrogen, as the administration suggests we can, why aren't we using using renewables for half our energy now!
"If you look at the code we believe has been copied in, it's not just a line or two, it's an entire section -- and in some cases, an entire program."
Can anyone make sense out of this statement? They are talking about code put into the kernel. How can multiple "entire program[s]" be copied into the kernel? Do they mean an entire.c files?
Well, search the kernel (2.4.18) for IBM copyrights. There is a ton of stuff in fs/jfs, s390 stuff, hotplug/cpqphp_* stuff.
Does any of this relate to high performance computing on multiple processors? I don't see that, but jfs is could be considered relevent to a high performance machine. It is possible the code in question was put in the kernel without an IBM copyright. I don't see anything remotely like a smoking gun
But if the FSF is right that OpenTV is violating the GPL, and if this behavior is found to be legal by the courts, the entire free-software and open-source movements could be derailed.
Derailed? No. Affected? Certainly.
The code that exists now can't be taken away, but the momentum can be stolen.
What if Dell were to sell hardware with DRM that would only run their modified, proprietary version of Linux. They would sell lots of these machines, because Linux is cool and useful and large companies have standing relationships with Dell. Companies that write and sell applications, particularly for business, will be targeting their code to run on Dell Linux, not Redhat. Linux code would have to be written to support either Linus's kernel or Dell's kernel; not both because Dell would want NDAs from developers.
The current momentum of increasing Linux support will be split into the free and proprietary streams. Free Linux would always have free gcc, emacs, X, and so on but the attention and money would go away. (Which I have no problem with.) Enough coders would be happy to program for the public domain to keep things going.
Nah, I'm just paranoid. But it's an amusing way to pass the time.
I should have called it a "gotcha", meaning it doesn't work in the way a reasonable person might assume. Sometimes called violating the Principle of Least Surprise. This is not to say that a program should be dumbed down to the point that it is instantly usable by any luser who can pilot a mouse. But if AFS transparently acts like the normal *nix FS when I use cp, mv, ls and so on, it is reasonable to assume that the protections also work as in a normal *nix FS.
I agree that "glitch" can imply "bug" which is not correct here. I am not sure how it is a "feature", though. I can see that it would be useful to enable stronger restrictions for file access via AFS, but not weaker.
Slashdot Mirror
Hear, hear! Watch someone using a computer, and see how time he wastes switching from keyboard to mouse. Especially for text searches. Old way: hitting C-f, finding the dialog with the mouse and getting to the text field, moving the dialog so you can see the page, and so on. New way: hit / and start typing. No mousing, no annoying dialog box. Amazing! Years of using "less" has conditioned me to use / for searches, and this works great.
"Type Ahead Find selects links preferentially"
This is an option you can change.
I will agree with the people who think he is propping up Linux. He is telling people that they don't have to go out and buy Sun or Microsoft software, because most of their free software is not under any threat. There is no need to get rid of all your Linux servers today, because you won't have to throw out your entire system if SCO should win; the worst case is that you have to get a different more or less equivalent kernel.
I think this is a voice in support in of Linux and free software for the ears of the PHBs.
That is why, under the US court system, you would never be sitting in such a jury. If this case went to a jury (I don't know if it could or not), no programmer or anyone with Linux or Unix experience would be allowed on the jury. You are guaranteed trial by a jury of the clueless.
You can't really distribute binaries compiled with gcc and claim to have a gnu-free system. Legally, yes. Philosophically, no.
When I recently reread Farhenheit 451, I caught a detail that I never paid attention to before. The firemen did not start burning books because the government wanted to eliminate them. The people demanded that thought-provoking, controversial and therefore disruptive works be destroyed. For any book, you could find some group who was bothered by it, so all books came under the kerosene.
I see aspects of a sort of reverse-Orwellian society today. Varied viewpoints and honest criticism may be available, but most people don't want to hear them, any more than Bradbury's society wanted their books. Given the choice between the happy myth and reality, people will choose the myth. How many Americans care about the truth of Iraqi WMD's, the Lynch "rescue", or whether Iraq was involved in the 9/11 attacks? Or that fuel cells are not an energy source? Far too few for my comfort.
This is worse than 1984, which envisions domination under an authoritarian government (as I remember; it has been a while). You don't have to beat down the people. Just tell the people the lies they want to hear, and they will do the rest. Whatever contradicts what people already want to believe will be ignored.
This is something fascinating about information access and the internet. The net does not serve to widely disseminate information, except in the most literal geographic sense. Instead, it allows people to form communities with others who already share the same opinions. Memes bounce around in a mostly closed community, building up power and credibility.
I can think of one concrete example. I received a forwarded email in 2000 of stupid statements allegedly made by Al Gore. I replied with an email from 1992 with the same set of quotes, but attributed to Dan Quayle. Did the original sender feel humiliated and send an apologetic retraction to everyone he had forwarded the message to? Of course not. The truth was easily available, but they liked the lie better.
What am I missing? The problem is with DNS. Use the numerical IP address of the webserver or mail server, and communications should proceed normally.
Right?
I agree completely. As an aside, when you want this kind of versioning, you can get emacs to write out versions as somestuff.~1~, somestuff.~2~,...with the most recent copy named just somestuff. The biggest reason I don't use this is because I can't get ls to sort them the way I would like, especially when it gets to version ~10~.
I can see it now... comments in both code misspelled it's for it and loose for lose, used fsck as an expletive, and referred to some mysterious substance known as pr0n.
IBM will then subpoena 600,000 registered /. users to prove that geeks can't spell.
It is not a fallacy to say that a source which provides erroneous information should not be relied upon.
I can see a fallacy if I used the fact of one false statement to conclude that other statements by her are probably false. It is however appropriate to say that I cannot a priori assume that her other statements are either true or false. Normally, you assume what someone says is true, especially if they claim to be an expert; Didio's clear misunderstanding of the GPL means that she no longer deserves to get the "benefit of the doubt." Her other statements must be evaluated on their own merit, just like any random /. post, rather than regarded as informed, expert opinion.
Didio: "Check the GPL and look at Section 0. It reads that the legal copyright holder of the source code has to explicitly put an assignment and copyright transfer notice into the beginning of the GPL. There is no concept of accidentally giving away the code to the GPL."
That's not what my copy of the GPL says. It says "This License applies to any program or other work which contains a notice placed by the copyright holder saying it may be distributed under the terms of this General Public License."
Further copying and modification are separate. The original author of the program placed the program under the GPL, and it and all distributed derivative works are likewise under the GPL.
If someone modifies and distributes the code, they don't in any way have to reaffirm the GPL notice. They certainly are not required to add additional copyright notices.
Didio's comment about the GPL is a complete misrepresentation, and she reveals herself here as an SCO pawn. I have no reason to believe anything else she said.
If Apple does this, they can sell their OS as an actual Unix. This would seem to be great as an advertising angle. It is one thing to know that an Apple machine would make a good server because it is Unix-like underneath, but it would have a great impact if it could say Unix right on the box. People who need a new Unix server may start thinking of Apple first.
They have already switched over to a Unix-y system, so why not make the most of it?
They cannot do this without throwing away their veil of secrecy.
I have source to at least 7 different kernel versions sitting around, mostly old RH cds. They would need to tell me which ones they want licensing fees for. And if their code is not compiled into the base kernel for all possible kernel configurations, then they would need to tell me what combinations of options cause their code to be included in the kernel. After all, if it is only in say jfs and smp code, I won't have it in my kernel, I won't be using it, and I won't owe them any licensing fee.
This is one difference open source makes. If the code were in a commercial product, they could announce which versions infringe. They would need to be much more specific here.
If they do indeed give out the specific information of what kernel versions and what modules infringe, it would be trivial to ferret out the offending code and be done with them.
This statement is true, but I would add that some supposedly infringing code could be proven to not come from SCO.
If SCO says blahblah.c is from their IP, and blahblah.c can be shown to have been lifted from BSD or to have grown from 50 different patches submitted to lkml by 50 different people, then their claim is demonstrably false.
If blahblah.c was mailed whole to Linus from someone at IBM in the right time period, only then would a comparison with SCO's source be necessary.
If they showed the part of the kernel code that they thought infringed, and it appeared even just a little bit possible that they were in the right, I think a big chunk of geek public opinion would swing their way. Why? Because the karma of the whole open source/free software community is at stake here, and people would not take kindly to IBM fscking up that karma.
I see your point, but I think this is too simplistic and not completely accurate.
I have seen people spend hours of confusion trying to figure out how to do something in MSWord, or some power point thingy, or Eudora. They didn't switch to different software.
There is a clear distinction between usability and marketability. Obviously commercial software has to be made marketable. My point is that so-called usability studies are often really marketability studies. This confusion prevents people from realizing that they can work to make the product both more usable and more marketable. The mistake is to unnecessarily sacrifice usability for marketability.
An obvious outgrowth of this thinking is programs with a sort of beginner mode and an expert mode. This is not uncommon under the *nix philosophy, because the gui and the guts are well separated and configurability is a goal. Any user can use the GUI, but there are command-line modes or ways of turning menus off.
Why should I care whether you can walk up to my machine and use it easily? (That's why I use xlock in the first place.)
Why do people concerned about usability spend so much time worrying about what people can do in the first 5 minutes after they sit down at a 'puter and so little time worrying about what people will be able to accomplish over the next five years?
Once when I was in a grocery store with a friend who spoke Russian, he overheard a conversation two recent arrivals were having. They were annoyed and complaining about the ice cream, because there were too many flavors to choose from.
It seems like the line between enough choice and too much choice isn't as concrete and obvious as some would think.
I need multiple virtual screens. Other than that, I probably could get most of the functionality I want out of twm with a well-written .twmrc. Looking at the man page, there are a lot of useful functions that aren't available unless you customize. For example, I use the fvwm equivalent of TwmWindows frequently; I didn't realize TwmWindows existed because you can't get to it in the default configuration.
And we had to chew the holes into them with our teeth.
I have, more than once, been incredibly relieved to find twm installed as a part of X on machines (not OS 10). Because when the install fails without getting all 90 billion parts of gnome or kde installed correctly, or using an old machine that can't handle the latest and greatest, I can use twm as a marginally useful window manager to start getting things done.
And when this happens, the one of the first things I do is download and install fvwm. Woohoo!
Consider that any expert that sees the "secret evidence" may then be unable to do any future work on the Linux kernel. This means the expert almost certainly cannot be someone familiar with the Linux kernel.
I think that this is SCO's secret weapon (in their minds). If the public knew what the supposed violations were, Linus, Eric Raymond, Bruce Perens, the FSF, and hundreds of others would be submitting amicus curiae briefs blowing SCO's case out of the water. They absolutely need to keep any reliable expert away from their "evidence".
If they accomplish this, then their lawyers and captive "experts" will try to dizzy the judge with techno-BS and win through Proof by Confusion. This is the only reason I can conceive of for keeping secret the location of the alleged violations.
Here's an idle speculation: They know their claims would be slaughtered if exposed to public scrutiny. Dozens of programmers could prove that the material wasn't taken from UNIX. But maybe they have an "expert" who can make a plausible case to a judge. To analyze the code for the court case, any Linux developer would have to sign an NDA essentially prohibiting her from ever working on the kernel again. And SCO thinks no developer with enough knowledge to disprove their claims would be willing to make the sacrifice.
It's a funny second-order trade secret argument. All of the code is public, but somehow they want the right to keep secret the knowledge of which parts of it used to be secret.
What I found on the web says that a car moving at highway speeds uses about 15 kW of power. The standard estimate for domestic power use is 1 kW averaged throughout the day.
Back of the envelope, let's say 10 million Californicators spend an hour a day in their cars. Averaged over 24 hours, this is over 6 GW. Entire daytime power usage in CA is about 35 GW (depending on season). And this doesn't account for SUVs using more power or commercial trucking.
I would be interested in seeing a real estimate, but it looks like this would require a substantial increase in power production facilities.
And this leads to a sticky question. If we can provide electricity via renewables to generate hydrogen, as the administration suggests we can, why aren't we using using renewables for half our energy now!
Can anyone make sense out of this statement? They are talking about code put into the kernel. How can multiple "entire program[s]" be copied into the kernel? Do they mean an entire .c files?
Well, search the kernel (2.4.18) for IBM copyrights. There is a ton of stuff in fs/jfs, s390 stuff, hotplug/cpqphp_* stuff.
Does any of this relate to high performance computing on multiple processors? I don't see that, but jfs is could be considered relevent to a high performance machine. It is possible the code in question was put in the kernel without an IBM copyright. I don't see anything remotely like a smoking gun
Derailed? No. Affected? Certainly.
The code that exists now can't be taken away, but the momentum can be stolen.
What if Dell were to sell hardware with DRM that would only run their modified, proprietary version of Linux. They would sell lots of these machines, because Linux is cool and useful and large companies have standing relationships with Dell. Companies that write and sell applications, particularly for business, will be targeting their code to run on Dell Linux, not Redhat. Linux code would have to be written to support either Linus's kernel or Dell's kernel; not both because Dell would want NDAs from developers.
The current momentum of increasing Linux support will be split into the free and proprietary streams. Free Linux would always have free gcc, emacs, X, and so on but the attention and money would go away. (Which I have no problem with.) Enough coders would be happy to program for the public domain to keep things going.
Nah, I'm just paranoid. But it's an amusing way to pass the time.
I agree that "glitch" can imply "bug" which is not correct here. I am not sure how it is a "feature", though. I can see that it would be useful to enable stronger restrictions for file access via AFS, but not weaker.