Just a question here, is there a Linux distro out there that currently has a "Dual Boot" option when you pop it into the CD-ROM drive that will reconfigure your Win2k/XP system as a dual boot system? I think a lot of the reluctance (even among hobbyists) to try Linux on their machines stems from the fear of losing Windows, and all the software you're used to using. Imagine a Linux CD you could pop in and click "Add Linux", which would reconfigure your HD with a Linux partition and allow your to choose from the bootloader. Complete risk-free Linux trial.
You could truly pass something like that around just like an AOL CD. And people wouldn't have to be afraid of it.
The hardware companies are eager to cut corners wherever possible to save a buck in producing their increasingly shotty, slipshod products. In the end, the removal of recovery media in favour of "recovery partitions" conveniently eating away at the consumer's free hard drive space is just another way of putting more cash into the company's greedy coffers. This has nothing to do with Microsoft's licencing or Windows XP or anything like that.
You can rest assured that, even if they were shipping Linux on these machines, they would probably still opt for providing as little actual installation media as they can get away with. Gateway has always tried to take advantage of consumer ignorance to push their below-average workmanship, which is why they're slowly slipping down the tubes. All the more reason to buy a decent system from Dell, or even better, Apple. You get what you pay for.
I was trying to decode this, but was having trouble with it until I figured out that it is in base64 encoding, not uuencode (as it appeared at first). If your Linux or Unix distribution does not have base64 installed by default, you can get it at http://www.fourmilab.ch/webtools/base64/. Thank you, Fair Use Guy, for promoting this tool.
In terms of technology, there is a big difference between a vinyl record and a CD.
In terms of technology, yes. In terms of the result of that technology, there's hardly any difference at all. Both systems translate a flat, one-dimensional stream of physical patterns into a flat, one-dimensional stream of audio. Neither has any support for programming, soundtrack selection, or any of the other advanced DVD features I was talking about.
Oh wait. Stupid comparison. There's no difference between a vinyl record and an audio CD. They're the same exact thing. Therefore, there is no difference between a VHS tape and DVD disk. The fact that its one generation newer in technology doesn't make it a fundamentally different thing.
Excuse me, but I never attempted to compare CDs to vinyl or DVDs to VHS. My post was comparing CDs to DVDs. Please read it again, bearing this in mind. But then, as you said, "this is a perfect example of what's wrong with higher education. A complete lack of critical thinking."
"Technologies naturally evolve over time, presumably getting better, and usually cheaper."
Yes, and the newer, better technologies are always initially more expensive than the common technologies that are already in place. DVD media will eventually drop in price as the technology is more refined and VHS is gradually deprecated and eliminated from the marketplace. You don't complain that the new 2.2GHz Pentium IVs are more expensive than the older, slower models, do you? Why should DVD media be any different?
I can do that with my "software" CDs so why shouldn't I do it with my "software" DVDs?
Beyond the cosmetic similarities, there's a world of difference between your CD-audio disc and your DVD-video movie. An audio CD is little more than a flat, 16-bit 44kHz stereo audio stream. Aside from easy track markings, a CD audio stream isn't much more flexible than a cassette or vinyl record, and because of this an audio CD hardly qualifies as software.
A DVD movie, on the other hand, is far more complex. The user can pick different soundtracks, different cuts, and all in all give more control to the user. This requires a significant amount of programming information to be present in the VOB files on the disc, much more than the jejune TOC system on an audio CD. There's a world of difference between the two media, and it's a fallacy to try to compare them.
DVD's are much more flexible and have far more capabilities than VHS or 35mm film or other old analogue media. Your average DVD disc generally has up to 3 times as much content as the equivalent VHS copy, what with different soundtracks, bonus footage, subtitles and captions in numerous languages, and all sorts of other extra features put onto the aluminium disk. These features are purely the work of software; the software logic on a DVD is far more complex than a VHS drive motor or film projector, and should be valued as such.
I for one will gladly pay twice as much for DVD content as I would for equivalent VHS content; the extensive capabilities of the DVD format make the medium that much more valuable. By declaring this practise illegal, I'm afraid Australia could be severely stifling the incentive of movie companies to include special DVD-only features. This move doesn't benefit anyone; both the content producers and the consumers suffer.
I too looked forward to the day my iBook came when I ordered. I would no longer have to sift through the duality of Linux's raw power and Windows' ubiquity and application support: I would have the best of both worlds! When I first plugged the iBook in and eagerly opened Terminal.app, I was pleased to be greeted by the lovely '%' tcsh prompt. 'ls' worked. 'du' worked. 'ifconfig' worked. Even "find . -name *.c -exec grep -l '{}' ';'" worked!
But as I dug deeper, the initial euphoria wore off. While MacOS X gives the impression of being a Unix, as one digs deeper into the system, it becomes more and more clear that it is anything but Unix:
No/etc. Well, technically, there is a/etc, but it is incredibly empty compared to what you may be used to in FreeBSD or Linux. None of the system's configuration is included in standard POSIX text files; Apple has opted to move everything into what they call a "NetInfo" registry. This is awful, to say the least. Obviously, Apple has failed to learn anything from the problems Microsoft's reliance on a central registry have propagated, and it wasn't long before I ended up having to reinstall MacOS X due to a corruption of the NetInfo database.
Not everything is a file. MacOS X violates this essential tenet of the Unix interface by hiding the implementation of several core functions such as keyboard and mouse I/O behind a "CoreGraphics" ObjC library. There is no/dev/mouse or/dev/kbd or any of the easy-to-use device nodes I've grown accustomed to coding for on Linux.
Everything above the Unix layer is proprietary. The Unix world on MacOS X is completely separate from the GUI world. The Unix directories are completely hidden from the Finder, and likewise one cannot start GUI apps from the console. There is none of the tight CLI-GUI integration seen in AmigaOS or BeOS or even Windows.
If you're looking for Unix, I suggest you make room on your iBook's drive for Linux, which runs very well on Mac hardware, and is much faster than MacOS X to boot. I'm afraid all of MacOS X's vaunted Unixness is little more than FUD.
That you all can take the time to petition for stupid TV shows to keep them from being cancelled, yet most of you can't seem to be arsed to write your representatives whenever terrible bills such as the PATRIOT act, DMCA, and SSSCA are in the process of being made into law. I would say our freedom is a bit more important than a stupid television show.
Don't get me wrong, I absolutely loved the show, but I always felt uncomfortable about its placement on Nickelodeon, a kid's network if there ever was one. The show's content, much like the class "Ren and Stimpy" episodes back from when Jim K. still had control of the show, is just too subversive for little kids. The show is more akin to things like "Family Guy" and "South Park" than it is to "Wild Thornberries" or whatever inane kiddie trash Klasky-Csupo is barfing over Nickelodeon these days. Hopefully the show will find a new, better home on Comedy Central or another such station where it would be a much better fit.
Throughout the years, Intel has proven itself as being almost as strong a backer of open-source software and Linux in particular as IBM. They have secured my business and my company's business through releasing their excellent P4-optimised compilers for Linux, as well as porting the operating system to their next-generation Itanium architecture.
While AMD processors may be slightly cheaper and run legacy x86 programs more quickly, SSE-optimised code compiled with Intel's compiler completely creams even the new Athlon XP "1900". Intel doesn't need to make up marketing numbers in order to make their processors look faster than they actually are. Our web server used to run on a 1.4GHz Thunderbird, which was cheap but notoriously unstable. And when the fan blew out a month ago, the whole computer was taken with it! Meanwhile, the Itanium system we used to replace it has been humming along without a hitch since it arrived. I do keep an Athlon box around for the occasional Quake 3 match, but for serious business, I support Intel and their strong open-source backing.
Well, on one hand, I have gotten some of the best, friendliest, most complete help with Linux from people on IRC and newsgroups. On the other hand, I have also received far more "RTFM" and other condescending, insulting replies than I'd care to. Microsoft technical support may be lacking in a number of ways, but you can bet the people on the other side of the phone want to keep their jobs and are generally reasonable, if a bit tedious to deal with. Linux support, like all things open-source, is the proverbial box of chocolates; depending on the time of day, phase of the moon, and the distribution of quantum fields in the universe, you could end up talking to a genuinely helpful fellow or some fucked-up adolescent douchebag who has a severe case of Schadenfreude.
Professional Linux support provided by Redhat et al. is OK, but there is the aspect of having to pay extra for the boxed Redhat distributions that include support. Your average poor CS undergrad isn't going to pay $59 for "free software", and is more likely to either stick with what he knows (Windows) or ask for Linux help in public fora such as IRC or Slashdot (and potentially be verbally abused). This is where the perception of Microsoft's superior technical support comes from.
If they can turn sodium(21) into magnesium(22), they're only two steps away from transforming Lead(82) into Gold(79)! Go Canada!:-)
A need for Distributed Content Storage
on
Adcritic Shuts Down
·
· Score: 5, Insightful
This is just one example of why peer-to-peer distributed systems like the Freenet project need to be developed. The Web is limited in that there needs to be somebody willing to maintain and update the servers on which data is stored, and that when a huge central resource like this can no longer afford to maintain their service, gigabytes of data can be potentially lost forever. What we need is a distributed system, where content is automatically propagated between nodes and can be downloaded from any one node, independent of venture capital and ad revenues.
Freenet does much of this, but still falls short of the ideal and still needs a lot of work to become viable. "Unpopular" data on Freenet is automatically destroyed to make room for more popular data, which makes it unsuitable for prolonged archival. There still isn't a decent search engine; finding data requires that you obtain the "key" from somebody who knows where to find it, which is inefficent and makes it hard for new Freenet users to locate data. If Freenet data could be made more permanent and easily searchable, or if somebody else could develop and promote a P2P network that isn't just a haven for warez and stolen music, it would become a great alternative to the struggling Web.
Above all, it has proven that you can burn through investor money in record time. Besides, who cares about money? The very existence of free software proves that quality can be produced without monetary incentives.
You don't need money to induce people to do what they want to do anyway.
But what of the grunt-work programming people don't want to do anyway? The most successful open-source projects are those that are interesting to work on -- Linux and FreeBSD are successful because hackers like working on operating systems, Crystal Space the 3D engine is successful because the engine is fun to work with. Your average free-software hacker, however, has several deficiencies: a lack of artistic talent among them. Why else do KDE and GNOME look more or less like wholesale ripoffs of CDE, Windows, and MacOS? Why are there no free *games* written on top of the Crystal Space engine? Hackers like to just hack -- they don't like to design or plan, preferring to "evolve" the software. While this is excellent for the kernel and other low-level programming, graphical interfaces and graphical programs need design, and this design generally requires billions of dollars worth of research and development work to develop.
Apple has it right -- open-source the plumbing of the operating system, the stuff that's available for free anyway, and keep the highly-developed GUI stuff proprietary. This synergy between the open and closed worlds is the best model I can think of. Free software needs somebody to feed design and R&D money into it because the hackers won't do that "boring" stuff for free.
All right, I'm sure your motivations are pure and all, but the fact that you're doing this under the GNU name is simply bound to get Stallman and his anal-retentive ways tangled into this mess. I have nothing against the GPL or the BSD licence (though I find BSD to be a tad too free, if you know what I mean), but GNU projects have a history of GPL bigotry and "our way or the highway" elitism which has caused no end of grief for many open-source projects (KDE, anyone?) and turned away many parties interested in Linux. From the poorly-written Slashdot article, I got the impression you were in the business of replacing non-compliant documentation, not merely cataloguing it as you say you are doing.
While I can appreciate your idealism, the "outdated" corporate model has proven to make money now. Of all the open-source companies, only Redhat and Apple are making a profit; Redhat by essentially having a monopoly on enterprise-level service and support for Linux, and Apple by receiving the majority of their profits from hardware. Free software will still be around, yes, but it will no longer be able to improve at the exponential rate it has been without corporate backing and most importantly money going into it. The Gordian knot of licencing in the free software world only serves to stroke the egos of RMS and the other zealots, while chasing away the money and support that could help bring their vision of "world domination" to life. People have been predicting the death of Microsoft and closed-source software since the inception of the GNU project, and yet their business model still is firmly entrenched because it works.
Here we have a license problem (you read the Debian story did you? Hmm? Yes I know that's slashdot:-) I didn't make it up. It is really a problem. Debian was going to move most of the LDP documents out of its main tree. We did everything in a hurry and it's now (mostly) fixed. But *prevention* is better - I'd rather have avoided this problem altogether!
The point I was trying to make is that this sort of thing shouldn't be a fucking problem in the first place. Debian has obviously survived for the past five years without nagging the LDP to make silly little fixes to their licences; it's positively atrocious of them to go "Change your licence or we're dumping your documentation; you have 48 hours to decide". A problem as pointless and minuscule as this should not be a big enough deal to make the Slashdot frontpage twice, nor should it require what amounts to an ultimatum to solve.
Also, licences don't sue people. I said this in my parent post, but it must be reiterated. An open-source licence is not so much a diction of what can and can't be done with the code as it is a statement of the author's intentions. I seriously doubt the Linux Documentation Project is going to call the Debian Linux distribution on some technicality of the GPL or DFSG. What could have been solved peacefully in a relatively quiet way by friendly parties shouldn't explode into an ultimatum situation and the discarding of reams of perfectly good manuals and HOWTOs only to be rewritten half-assed.
Companies are acclimated to the closed-source view, and bother the producers and consumers of software tools are usually willing to bargain or compromise for special licencing provisions and/or rights to source code. The average open-source hacker is far more idealistic, and will ravenously stick to his choice of licence (just look at all the GPL vs. BSD flamewars if you don't believe me).
There is also the trouble of actually finding the authors who hold the copyright to the code; there is a TON of old, unmaintained software that has outdated e-mail addresses and no other way of contacting the author. And for large collaborative projects like Linux, copyright is split between the thousands of contributors that have added to the project over its lifespan. A corporation provides a single, monolithic entity to approach for licencing; an open-source project provides an unkempt mishmash of hundreds of hard-to-find developers with different ideals and personalities; your average company isn't going to bother rectifying licence terms with that many different, unpredictable people.
All this bickering over licence is ludicrously counterproductive. Licences don't sue people, people do. I hardly think the writer of a GPL piece of software would care what you do with the accompanying documentation, and it is baffling to me why the Debian nuts think documentation needs to be under its own special licence instead of using the GPL along with the software it documents. The reason I switched to open-source software was to get away from all the stupid EULA politics and policing of the traditional software world. I hate to see this pigheadedness seep into the open-source world.
Writing documentation is an incredibly difficult task, and few people do it well; to throw out an incredibly useful and well-written resource simply because of a miniscule licencing technicality is both horribly naïve and terribly anal behaviour. How does this guy think he'll be able to rewrite, say, all the Linux man pages without (a) having the original manpages as a reference and (b) quite possibly not being anywhere near as good a documentor as the original Linux Documentation Project? Open-source documentation is scarce and hard to come by as it is, why does Debian feel the need to exacerbate this shortcoming even further?
This hole only works if the browser-shell integration "feature" of IE >4.0 is enabled. This is easy to disable, if you happen to have a Windows 95 CD on hand:
Copy your current explorer.exe, shell32.dll, comdlg32.dll, notepad.exe and wordpad.exe to a backup location in case things go haywire. (I've done this before on Windows 98 and ME boxes without problems, but it's always good to be safe).
Insert the Windows 95 CD, and start a dos prompt.
From the prompt, enter:
d: (or whatever your CD drive is)
cd win95
extract/a/l c:\your\windows\desktop win95_02.cab comdlg32.dll explorer.exe shell32.dll notepad.exe wordpad.exe
You should have the files listed above appear on your desktop. Now shut down into DOS mode, and copy the new shell32.dll and comdlg32.dll into your Windows SYSTEM directory, and copy explorer.exe, notepad.exe and wordpad.exe into your WINDOWS directory, and reboot Windows. (If you're using ME, you can go into c:\windows\system.ini and change your shell to taskman.exe in order to be able to replace explorer and the other system files)
Your system should come up with the old Windows 95 shell, which doesn't have any of the IE integration bullshit. IE will still launch as a separate application (with an Office-style splash screen, even!) and since the IE dll's aren't stuck in your memory all the time, your system should be a bit faster too.
Of course, after doing this, the next step is to replaceyourbrowser, but that goes without saying.:-)
I'm sure I'm not alone when I ask you, CmdrTaco, why did you stop posting quickies for so long? They are an essential part of the Slashdot experience, a completely silly and mindless break from all the serious, downbeat Microsoft and DMCA stories that make up the majority of Slashdot's articles. Thank you for bringing them back, but why did you stop posting them in the first place?
... but it bugs me that nobody who marked you +5 seemed to notice that you never mentioned what this feature was!!!
If you had bothered to actually read my post, you would see that I do tell you: fine-grained dependency declaration. A DEB package references the other packages it requires, not the libraries or binaries contained in the package, thereby alleviating the problem of depending on different library versioning conventions. It also allows for pre- and post-dependencies, which help apt handle the often complex task of updating between major versions of Debian, by making sure that packages are updated in the proper order. This way, apt can successfully upgrade you from the old, libc5-based 1.3 distro to the latest unstable branch with practically no intervention. Try upgrading from Redhat 5.x to 7.2 sometime, and see how much work you have to put into it.
Debian has the sense to both provide for people who want the latest and greatest toys to play with (in their "unstable" branch, which is generally more usable than the moniker suggests) and those who need a thoroughly tested and secured environment for production work (in their aptly-named "stable" branch). Whereas other distributors like Redhat and Mandrake find themselves forced to compromise between having the latest gimmicks and a stable base, Debian provides for both needs.
Debian unstable has all of the latest packages a mere "apt-get update && apt-get upgrade" away, including kernels 2.4.x, evolution, KDE 2.2.2, nautilus, etc. The main reason for the "unstable" label is that they don't support it as a release, but in my experience Debian "unstable" has always been stable enough for desktop use. (This is not to be confused with Debian's "testing" branch, which is, er, messy:-)
For servers and workstations that require high availability, though, you don't use the latest version of anything. Ever. Debian stable, while certainly not the bleeding edge, provides a set of software that has been proven by time and experience to work. And the latest security fixes are always available from security.debian.org. Debian stable is exactly what is needed for a production environment.
People who criticise Debian releases for being out-of-date need only look at the numerous security and stability issues Redhat and Mandrake have to suffer through in order to maintain currency. Debian provides a stable, proven base upon which you can easily upgrade to the latest "unstable" branch if you feel like you need the latest toys to mess with.
...not with the various hacks being piled on top of it to support its deficiencies. Apt works so well with Debian because the DEB archive format supplies incredibly detailed dependency and install-order data, so that apt-get's update feature can quickly and precisely determine what packages to download, and in what order to install them, without updating packages for which there is a dependency on an older version.
RPM, on the other hand, specifies extremely inadequate information to support such a tool without a lot of extra hacking. The RPM format at best only provides the name and major version of any dynamic libraries a package requires. Since different distributions and upstream authors all seem to have their own ideas on how to use dynamic library versioning, this quickly degenerates into the dependency hell that anybody who has tried to install or upgrade a reasonably complex program like GNOME on an RPM-based distro has probably encountered at least once in their experience.
Instead of dragging our feet with RPM and all its drawbacks, why not just move distributions over to dpkg/apt/DEB management like Debian, or FreeBSD-style ports? It is all free software, and there should be no problem with licences or any political bollocks. The rotten RPM format, which has somehow managed to become the Linux standard, is a monstrosity only beaten in kludginess by Windows' inept software management system, and with the MSI package format introduced with Windows 2000, even Microsoft is making RPM look crappy. Leave it behind, and move to a better system.
Slashdot Mirror
Just a question here, is there a Linux distro out there that currently has a "Dual Boot" option when you pop it into the CD-ROM drive that will reconfigure your Win2k/XP system as a dual boot system? I think a lot of the reluctance (even among hobbyists) to try Linux on their machines stems from the fear of losing Windows, and all the software you're used to using. Imagine a Linux CD you could pop in and click "Add Linux", which would reconfigure your HD with a Linux partition and allow your to choose from the bootloader. Complete risk-free Linux trial.
You could truly pass something like that around just like an AOL CD. And people wouldn't have to be afraid of it.
Does something like this already exist?
You can rest assured that, even if they were shipping Linux on these machines, they would probably still opt for providing as little actual installation media as they can get away with. Gateway has always tried to take advantage of consumer ignorance to push their below-average workmanship, which is why they're slowly slipping down the tubes. All the more reason to buy a decent system from Dell, or even better, Apple. You get what you pay for.
I was trying to decode this, but was having trouble with it until I figured out that it is in base64 encoding, not uuencode (as it appeared at first). If your Linux or Unix distribution does not have base64 installed by default, you can get it at http://www.fourmilab.ch/webtools/base64/. Thank you, Fair Use Guy, for promoting this tool.
A DVD movie, on the other hand, is far more complex. The user can pick different soundtracks, different cuts, and all in all give more control to the user. This requires a significant amount of programming information to be present in the VOB files on the disc, much more than the jejune TOC system on an audio CD. There's a world of difference between the two media, and it's a fallacy to try to compare them.
I for one will gladly pay twice as much for DVD content as I would for equivalent VHS content; the extensive capabilities of the DVD format make the medium that much more valuable. By declaring this practise illegal, I'm afraid Australia could be severely stifling the incentive of movie companies to include special DVD-only features. This move doesn't benefit anyone; both the content producers and the consumers suffer.
But as I dug deeper, the initial euphoria wore off. While MacOS X gives the impression of being a Unix, as one digs deeper into the system, it becomes more and more clear that it is anything but Unix:
- No
/etc. Well, technically, there is a /etc, but it is incredibly empty compared to what you may be used to in FreeBSD or Linux. None of the system's configuration is included in standard POSIX text files; Apple has opted to move everything into what they call a "NetInfo" registry. This is awful, to say the least. Obviously, Apple has failed to learn anything from the problems Microsoft's reliance on a central registry have propagated, and it wasn't long before I ended up having to reinstall MacOS X due to a corruption of the NetInfo database.
- Not everything is a file. MacOS X violates this essential tenet of the Unix interface by hiding the implementation of several core functions such as keyboard and mouse I/O behind a "CoreGraphics" ObjC library. There is no
/dev/mouse or /dev/kbd or any of the easy-to-use device nodes I've grown accustomed to coding for on Linux.
- Everything above the Unix layer is proprietary. The Unix world on MacOS X is completely separate from the GUI world. The Unix directories are completely hidden from the Finder, and likewise one cannot start GUI apps from the console. There is none of the tight CLI-GUI integration seen in AmigaOS or BeOS or even Windows.
If you're looking for Unix, I suggest you make room on your iBook's drive for Linux, which runs very well on Mac hardware, and is much faster than MacOS X to boot. I'm afraid all of MacOS X's vaunted Unixness is little more than FUD.That you all can take the time to petition for stupid TV shows to keep them from being cancelled, yet most of you can't seem to be arsed to write your representatives whenever terrible bills such as the PATRIOT act, DMCA, and SSSCA are in the process of being made into law. I would say our freedom is a bit more important than a stupid television show.
Don't get me wrong, I absolutely loved the show, but I always felt uncomfortable about its placement on Nickelodeon, a kid's network if there ever was one. The show's content, much like the class "Ren and Stimpy" episodes back from when Jim K. still had control of the show, is just too subversive for little kids. The show is more akin to things like "Family Guy" and "South Park" than it is to "Wild Thornberries" or whatever inane kiddie trash Klasky-Csupo is barfing over Nickelodeon these days. Hopefully the show will find a new, better home on Comedy Central or another such station where it would be a much better fit.
While AMD processors may be slightly cheaper and run legacy x86 programs more quickly, SSE-optimised code compiled with Intel's compiler completely creams even the new Athlon XP "1900". Intel doesn't need to make up marketing numbers in order to make their processors look faster than they actually are. Our web server used to run on a 1.4GHz Thunderbird, which was cheap but notoriously unstable. And when the fan blew out a month ago, the whole computer was taken with it! Meanwhile, the Itanium system we used to replace it has been humming along without a hitch since it arrived. I do keep an Athlon box around for the occasional Quake 3 match, but for serious business, I support Intel and their strong open-source backing.
Professional Linux support provided by Redhat et al. is OK, but there is the aspect of having to pay extra for the boxed Redhat distributions that include support. Your average poor CS undergrad isn't going to pay $59 for "free software", and is more likely to either stick with what he knows (Windows) or ask for Linux help in public fora such as IRC or Slashdot (and potentially be verbally abused). This is where the perception of Microsoft's superior technical support comes from.
If they can turn sodium(21) into magnesium(22), they're only two steps away from transforming Lead(82) into Gold(79)! Go Canada! :-)
Freenet does much of this, but still falls short of the ideal and still needs a lot of work to become viable. "Unpopular" data on Freenet is automatically destroyed to make room for more popular data, which makes it unsuitable for prolonged archival. There still isn't a decent search engine; finding data requires that you obtain the "key" from somebody who knows where to find it, which is inefficent and makes it hard for new Freenet users to locate data. If Freenet data could be made more permanent and easily searchable, or if somebody else could develop and promote a P2P network that isn't just a haven for warez and stolen music, it would become a great alternative to the struggling Web.
Yes, poorly-documented, half-finished, unorganised quality.
You don't need money to induce people to do what they want to do anyway.
But what of the grunt-work programming people don't want to do anyway? The most successful open-source projects are those that are interesting to work on -- Linux and FreeBSD are successful because hackers like working on operating systems, Crystal Space the 3D engine is successful because the engine is fun to work with. Your average free-software hacker, however, has several deficiencies: a lack of artistic talent among them. Why else do KDE and GNOME look more or less like wholesale ripoffs of CDE, Windows, and MacOS? Why are there no free *games* written on top of the Crystal Space engine? Hackers like to just hack -- they don't like to design or plan, preferring to "evolve" the software. While this is excellent for the kernel and other low-level programming, graphical interfaces and graphical programs need design, and this design generally requires billions of dollars worth of research and development work to develop.
Apple has it right -- open-source the plumbing of the operating system, the stuff that's available for free anyway, and keep the highly-developed GUI stuff proprietary. This synergy between the open and closed worlds is the best model I can think of. Free software needs somebody to feed design and R&D money into it because the hackers won't do that "boring" stuff for free.
All right, I'm sure your motivations are pure and all, but the fact that you're doing this under the GNU name is simply bound to get Stallman and his anal-retentive ways tangled into this mess. I have nothing against the GPL or the BSD licence (though I find BSD to be a tad too free, if you know what I mean), but GNU projects have a history of GPL bigotry and "our way or the highway" elitism which has caused no end of grief for many open-source projects (KDE, anyone?) and turned away many parties interested in Linux. From the poorly-written Slashdot article, I got the impression you were in the business of replacing non-compliant documentation, not merely cataloguing it as you say you are doing.
While I can appreciate your idealism, the "outdated" corporate model has proven to make money now. Of all the open-source companies, only Redhat and Apple are making a profit; Redhat by essentially having a monopoly on enterprise-level service and support for Linux, and Apple by receiving the majority of their profits from hardware. Free software will still be around, yes, but it will no longer be able to improve at the exponential rate it has been without corporate backing and most importantly money going into it. The Gordian knot of licencing in the free software world only serves to stroke the egos of RMS and the other zealots, while chasing away the money and support that could help bring their vision of "world domination" to life. People have been predicting the death of Microsoft and closed-source software since the inception of the GNU project, and yet their business model still is firmly entrenched because it works.
The point I was trying to make is that this sort of thing shouldn't be a fucking problem in the first place. Debian has obviously survived for the past five years without nagging the LDP to make silly little fixes to their licences; it's positively atrocious of them to go "Change your licence or we're dumping your documentation; you have 48 hours to decide". A problem as pointless and minuscule as this should not be a big enough deal to make the Slashdot frontpage twice, nor should it require what amounts to an ultimatum to solve.
Also, licences don't sue people. I said this in my parent post, but it must be reiterated. An open-source licence is not so much a diction of what can and can't be done with the code as it is a statement of the author's intentions. I seriously doubt the Linux Documentation Project is going to call the Debian Linux distribution on some technicality of the GPL or DFSG. What could have been solved peacefully in a relatively quiet way by friendly parties shouldn't explode into an ultimatum situation and the discarding of reams of perfectly good manuals and HOWTOs only to be rewritten half-assed.
There is also the trouble of actually finding the authors who hold the copyright to the code; there is a TON of old, unmaintained software that has outdated e-mail addresses and no other way of contacting the author. And for large collaborative projects like Linux, copyright is split between the thousands of contributors that have added to the project over its lifespan. A corporation provides a single, monolithic entity to approach for licencing; an open-source project provides an unkempt mishmash of hundreds of hard-to-find developers with different ideals and personalities; your average company isn't going to bother rectifying licence terms with that many different, unpredictable people.
Writing documentation is an incredibly difficult task, and few people do it well; to throw out an incredibly useful and well-written resource simply because of a miniscule licencing technicality is both horribly naïve and terribly anal behaviour. How does this guy think he'll be able to rewrite, say, all the Linux man pages without (a) having the original manpages as a reference and (b) quite possibly not being anywhere near as good a documentor as the original Linux Documentation Project? Open-source documentation is scarce and hard to come by as it is, why does Debian feel the need to exacerbate this shortcoming even further?
- Copy your current explorer.exe, shell32.dll, comdlg32.dll, notepad.exe and wordpad.exe to a backup location in case things go haywire. (I've done this before on Windows 98 and ME boxes without problems, but it's always good to be safe).
- Insert the Windows 95 CD, and start a dos prompt.
- From the prompt, enter:
- You should have the files listed above appear on your desktop. Now shut down into DOS mode, and copy the new shell32.dll and comdlg32.dll into your Windows SYSTEM directory, and copy explorer.exe, notepad.exe and wordpad.exe into your WINDOWS directory, and reboot Windows. (If you're using ME, you can go into c:\windows\system.ini and change your shell to taskman.exe in order to be able to replace explorer and the other system files)
Your system should come up with the old Windows 95 shell, which doesn't have any of the IE integration bullshit. IE will still launch as a separate application (with an Office-style splash screen, even!) and since the IE dll's aren't stuck in your memory all the time, your system should be a bit faster too.d: (or whatever your CD drive is) /a /l c:\your\windows\desktop win95_02.cab comdlg32.dll explorer.exe shell32.dll notepad.exe wordpad.exe
cd win95
extract
Of course, after doing this, the next step is to replace your browser, but that goes without saying. :-)
I'm sure I'm not alone when I ask you, CmdrTaco, why did you stop posting quickies for so long? They are an essential part of the Slashdot experience, a completely silly and mindless break from all the serious, downbeat Microsoft and DMCA stories that make up the majority of Slashdot's articles. Thank you for bringing them back, but why did you stop posting them in the first place?
If you had bothered to actually read my post, you would see that I do tell you: fine-grained dependency declaration. A DEB package references the other packages it requires, not the libraries or binaries contained in the package, thereby alleviating the problem of depending on different library versioning conventions. It also allows for pre- and post-dependencies, which help apt handle the often complex task of updating between major versions of Debian, by making sure that packages are updated in the proper order. This way, apt can successfully upgrade you from the old, libc5-based 1.3 distro to the latest unstable branch with practically no intervention. Try upgrading from Redhat 5.x to 7.2 sometime, and see how much work you have to put into it.
Debian unstable has all of the latest packages a mere "apt-get update && apt-get upgrade" away, including kernels 2.4.x, evolution, KDE 2.2.2, nautilus, etc. The main reason for the "unstable" label is that they don't support it as a release, but in my experience Debian "unstable" has always been stable enough for desktop use. (This is not to be confused with Debian's "testing" branch, which is, er, messy :-)
For servers and workstations that require high availability, though, you don't use the latest version of anything. Ever. Debian stable, while certainly not the bleeding edge, provides a set of software that has been proven by time and experience to work. And the latest security fixes are always available from security.debian.org. Debian stable is exactly what is needed for a production environment.
People who criticise Debian releases for being out-of-date need only look at the numerous security and stability issues Redhat and Mandrake have to suffer through in order to maintain currency. Debian provides a stable, proven base upon which you can easily upgrade to the latest "unstable" branch if you feel like you need the latest toys to mess with.
RPM, on the other hand, specifies extremely inadequate information to support such a tool without a lot of extra hacking. The RPM format at best only provides the name and major version of any dynamic libraries a package requires. Since different distributions and upstream authors all seem to have their own ideas on how to use dynamic library versioning, this quickly degenerates into the dependency hell that anybody who has tried to install or upgrade a reasonably complex program like GNOME on an RPM-based distro has probably encountered at least once in their experience.
Instead of dragging our feet with RPM and all its drawbacks, why not just move distributions over to dpkg/apt/DEB management like Debian, or FreeBSD-style ports? It is all free software, and there should be no problem with licences or any political bollocks. The rotten RPM format, which has somehow managed to become the Linux standard, is a monstrosity only beaten in kludginess by Windows' inept software management system, and with the MSI package format introduced with Windows 2000, even Microsoft is making RPM look crappy. Leave it behind, and move to a better system.