Well up to four months ago I was running a PII-400 with 128MB RAM and was running all the same stuff on Debian unstable with no problems at all. Granted, the bandwidth thing is an issue, but if you have broadband then you're fine. Dependencies weren't an issue there either. Again, most people don't need the latest software, and are perfectly happy re-using stuff for ages (witness yourself with Win95 of all things). You don't have to sit there with Redhat 7.0 and try and run the latest stuff, you can just use the older stuff. If you want the latest stuff, you might have to shell out for the hardware that's capable of running it. Your PII would handle it just fine (with a lightweight WM like icewm), and a Redhat CD set from cheapbytes or whatever wouldn't have cost you very much.
Redhat 7.0 was released on September 25th, 2000. That's ancient. Not only are Redhat's 7.0 releases notoriously buggy (and 7.0, iirc was really bad even for Redhat). I've not had a problem installing flash in years, and getting out of depenendancy hell would be no different on a windows box. If you want to use newer software like OO, you need to have newer libraries. You could have installed Debian and had all this stuff taken care of you for free, and I think Mandrake also. Even then, you can't expect the best on a P-233. Sorry, that's not even close to a modern desktop these days.
I'm not denying the steep learning curve of Linux at all, but you're going out of your way to make life difficult on yourself. You say using Redhat9 on a P-233 would be crazy? Well why would that be any more crazy than trying to install RH7 and run the software that comes with 9? That's exactly waht you're doing here. You can install RH9 (or Fedora these days) and just run a lightweight wm and you should be fine. You went about things the wrong way, assuming RH9 is different at some fundamental level from 7, which it's not.
This is the sort of thing that fascinates me. "What Linux needs to conquer the desktop" changes every year in the most astounding ways. A few years ago, it was "We need a browser as good as IE because Netscape is crap and Mozilla is slow bloated and clunky". Then it was "We need an Office replacement". Then it was "We need a good UI, because KDE and Gnome just aren't good enough".
Now it's getting so specialized. The complaints these days are "we need a real Outlook clone that's interoperable/groupware that's compatible with exchange" and that's about it. No more talk of corporate support. No more talk of good desktop UI's. No more talk of nice pretty installers. No more talk of hardware autodetection. No more talk of drivers. No more talk of good Office software, Web browsers, media players, or pretty widgets. I've got news for you, the train's been coming, slow but sure. All the old complaints are dying off and things keep moving forward. I wonder what people will say needs to happen before Linux conquers the desktop next?
What in the world are you using? LFS? Gentoo? Why are you compiling gaim? I've never compiled gaim myself and I've been a Linux and gaim user for five years now.
If you sit a user down at a windows box, you'll never see them say "I want to customize the UI of this thing, give me a different window manager now!" They'll just use what's there. In the case of linux, if it's KDE2 then they'll use that. If it's KDE3 then they'll use that. If it's fvwm then they'll likely have some trouble until you show them how to work it. My largely computer illiterate friends had no troubles at all with windowmaker or icewm.
And as for dependencies, use your distro properly! Debian, Redhat, Mandrake, SuSE, ad infitum will have programs to properly manage depenancies so you don't have to. This problem was solved ages ago. apt and RPM were written well before I started using Linux, so it's not like they haven't been around out in the open for you to find.
Sure, maybe this or that distro might not have everything perfectly set up the way you want it, but then again neither does windows initially. Things still have to be installed, and just because you might be more used to double clicking on some random.exe you got from download.com than selecting a program from a list in synaptic and telling it to install doesn't mean that you can't be retrained in all of 5 minutes. I sure did that with a friend of mine and he had no troubles installing the software that came with his distro. Learn to use the tools that are there for newbies and you'll be fine. It's not really so hard, and I can personally attest to this because the learning curve has dropped significantly since I started using Linux. It's ready for those who are willing to use it.
Homestar Runner is one of the best sites on the net. XiaoXiaoMovie is awesome as well. I can't think of any other sites like these off the top of my head, but I'm sure others can add to the list.
I personally hope Fedora actually gains some critical mass, but the problem with it seems like the same one that plagued mozilla at first. It started as a closed project, and it took a while to attract people to get that critical mass. With something like Debian or Gentoo, they grew out of openess to begin with, so by the time the project was big it already had a healthy mix of developers. Fedora will attract people, if it's really a worthwhile project, and if not then those people will move to Gentoo or Debian.
Just out of curiosity, which character in Watchmen would you be? And if you couldn't play Morpheus, who would you play in the Sandman mythos?
I'm no actor, but I'd probably go with Nightowl (the young one) from Watchmen and voicing Matthew the Raven from Sandman. The first has this great sense of being lost in the world, and I adore Matthew, especially for the way he handles things in The Wake.
It's available in experimental. It probably hasn't hit the general user populace (and really hasn't hit those outside of the Debian userbase) but experimental is slated to take on certain roles once held by unstable. Any sort of -snapshot packages (basically cvs pulls that are pre-release) now go in to experimental rather than unstable. Unstable is for packages that the maintainer considers release quality but hasn't tested widely. Experimental is for when you're still working on the thing, but would like to allow testing anyway.
At least, that's the plan. It's pretty new but it seems to be working so far. Hopefully it'll lead to faster releases.
Anyhow, the xfree 4.3 packages are in experimental. The TODO list for those packages is posted in the subversion archive where the Debian packaging development is done (you can get to it from the Debian X Strikeforce webpage) so you can see how much is left before the packages migrate to unstable. As it is though, the 4.3 packages are there if you want them.
I agree, that Debian has some problems with usability. While there are some things it does very well usability-wise, including the menu system, the alternatives system, the stock kernels, the kernel-package and patch system, and conffiles, it does lack cohesiveness in other ways. A big part of this is that the project is so damn big, it's basically impossible to wrangle anything so complex and difficult as usability guidelines (although I've considered trying to write some myself, but have no idea where to start).
On the other hand, Debian does provide options to do some things rather well. You can install discover and have hardware autodetection if you want it. You could have used PGI in the past to get yourself that graphical installer with no real fuss. The problem is the plethora of choices and sorting through them, not ignoring the users. I would very much like to see the usability of Debian extended (the new installer does have a gtk frontend, by the way) but the problem is, as usual, both related to politics and manpower.
Ultimately, I'd love to hear suggestions on ways that Debian could improve usability. I personally think it's a power user's dream come true in terms of usability. How would you improve it, beyond a graphical installer? Hardware autodetection is also, I believe, a standard component of the new installer, so don't list that please.
Not really. The mplayer situation has been progressing, especially because some of the flames have calmed down, and the mplayer developers have been a bit nicer on the Debian mailing lists. There are still major hurdles to clear, mainly social ones. Many Debian developers, including those that administer the ftp archive, don't trust the mplayer developers yet. Hopefully it'll get in at some point though.
having packages for non-free stuff is good, and NECESSARY, as well. Yes, anyone can make them.. but it really helps having a central repository. IN the real business world, you can only take the need for freedom so far. I love Debian servers, but at some point there is non-free stuff I *need* to run.
Well, apt-get.org has been acting as a good central routing point for unofficial packages for quite a while, and it works well. One thing that any business can do is use apt-get.org to find non-free stuff, and use those unofficial sources. They can also very easily set up their own private archive with the non-free packages they've downloaded to distribute amongst their own machines. You might not have the BTS, but perhaps that will motivate someone to set up a separate BTS and mailing list for the non-free development. Bruce has already mentioned non-free.org for this sort of thing, and I see no reason why it can't have its own BTS and such.
Yes, and it's about time too. I'm a little hesitant about the international users issue that kept coming up during the last discussion about removing non-free, but compared to the days of yore when we couldn't contemplate life without netscape, this is a small thing, and will probably fix itself when the need arises. I can almost see it as the project coming to fruition in a way.
Linux users are debating whether or not some text file is truly "free" when it is in a non-GNU directory or if RMS doesn't pee on it
Troll feeding time. Linux users aren't doing this, it's the developers. The same way Apple developers agonize over the exact shade of blue to use for their buttons (and rightly that they should) Linux developers agonize over this sort of thing. Honestly, it's one of the major strengths of Debian, that the developers spend their time wrangling over this and that detail so their users don't have to. Apple provides the exact same service by making their interface pleasant to look at so you don't have to go hunting for a better theme for your damn desktop. It's time saving, just for people interested in different things.
You would wish that some distro would have its social contract be "To produce a great distribution with the latest software that is stable". Succinct, and what it should be all about.
Good luck. The latest software isn't stable. It's got lots of hidden bugs. That's what the testing process is for. Different programmers test to different degrees. Obviously the linux kernel is heavily tested, but what about all the little programs that come with Gnome?
And then once the software is in the hands of the distro makers, they have to package it properly. Now, believe it or not, but packaging something well is tough. It's like programming, so you have to shake the bugs out of that too over time.
I guess it comes down to a different definition of stable. If you think all software is stable right out of the gate because it runs, then you've got a different definition of stable than most Debian developers. If you want that stuff, there's a version of Debian you can track: it's called unstable (and you can even sprinkle in experimental, if you really have the faith).
I really don't understand the general Linux user's need to get the latest and greatest at all times. Most of us ran windows for years, and simply waited for Microsoft to say "Ok, it's ready, here you go" every two or more years before upgrading. Debian's release schedule isn't a whole lot different, but you can simply see what's going on behind the scenes, so people tend to get impatient. How quickly we forget.
Because it goes against the very philosophy of Debian, right down to its initial charter if you want to go read it. Branden is one of the few people who actually still give a damn about what the project was originally founded for. He produces quality work (the Debian X packages are top notch, and he basically manages the porting work to other arch's because XFree puts out unportable crap) and he sticks to his ideals in the face of criticism.
He's not stopping you from using qmail, nor is he stopping people from putting up their own apt sources (people do it every day!) that host non-free, he just wants it out of the project so it can be free of this bizarre dichotomy.
What's really interesting here is that this moves them a little closer to the way the Gentoo people operate. Take a look at Their Social Contract for comparison purposes.
No. No, not at all. In fact it moves them farther away from Gentoo. In Gentoo's Social Contract, there is nothing explicitly stated about Documentation, but rather refers explicitly to software as binaries or sources. Debian has been working on productive discussions with the FSF over the GFDL for over two years, and this change is a direct result from those discussions. Most Debian Developers feel that documentation qualifies as software, and should be included under the DFSG as well.
Everything in Gentoo's Social contract is basically directly lifted from Debian's, although they decided not to take it all. The Gentoo people don't operate on nearly the same strict standards of Freedom that Debian does, and the differences in the Social Contracts, including the latest change, demonstrates that. If the Gentoo people decide to move in this direction too, it'll be because of more than two years of hard wrangling on debian-legal.
Historical reasons, apparently. When exim was chosen as the default MTA for Debian, postfix's license was in review and it wasn't sure as to whether or not it satisfied the DFSG. There was a discussion about changing the default MTA for the next release (here is the thread) but no real progress was made. exim4 seems like it'll be a perfectly fine choice, especially since no truly compelling argument could be made to switch, as much as I'd root for postfix myself.
Any serious CS student will find out about Linux anyway, and any serious CS student will try it out. If they like it, they'll keep using it, and if not, they'll go back to Windows. If they like it enough to contribute, then they'll learn to program it, and if they don't then they'll just program what they learned in class.
Now how is this any different than the way things already were?
The thing about apt, at least for Debian users, is that there is a strong and established mirror system already in place that handles the load just fine. I don't know how this applies to aptrpm users and whatnot.
As a sidenote, one of the main dpkg developers was playing around with using bittorrent in conjunction with the mirror system for a while, so the problem has at least been toyed with. I don't know if we'll ever see it in action though, since the need just isn't there right now.
Then what does developing the cygwin libraries make you?
He's talking about platforms. If you're developing on cygwin, you're developing for a UNIX platform that just happens to be running on top of Windows. It can be moved to a genuine UNIX platform with ease, so you're not actually tied to Windows. This is just like developing on Linux and then moving your app to BSD. So you're not a serf or a sharecropper, you're genuinely in control of what you're building.
Just because mass market PC products are made for windows doesn't mean that they won't run Linux just fine. You'd have a very difficult time demonstrating that Windows can run on more hardware configurations than Linux, especially given that Linux supports totally different architectures in addition to x86.
Slashdot Mirror
Well up to four months ago I was running a PII-400 with 128MB RAM and was running all the same stuff on Debian unstable with no problems at all. Granted, the bandwidth thing is an issue, but if you have broadband then you're fine. Dependencies weren't an issue there either. Again, most people don't need the latest software, and are perfectly happy re-using stuff for ages (witness yourself with Win95 of all things). You don't have to sit there with Redhat 7.0 and try and run the latest stuff, you can just use the older stuff. If you want the latest stuff, you might have to shell out for the hardware that's capable of running it. Your PII would handle it just fine (with a lightweight WM like icewm), and a Redhat CD set from cheapbytes or whatever wouldn't have cost you very much.
Redhat 7.0 was released on September 25th, 2000. That's ancient. Not only are Redhat's 7.0 releases notoriously buggy (and 7.0, iirc was really bad even for Redhat). I've not had a problem installing flash in years, and getting out of depenendancy hell would be no different on a windows box. If you want to use newer software like OO, you need to have newer libraries. You could have installed Debian and had all this stuff taken care of you for free, and I think Mandrake also. Even then, you can't expect the best on a P-233. Sorry, that's not even close to a modern desktop these days.
I'm not denying the steep learning curve of Linux at all, but you're going out of your way to make life difficult on yourself. You say using Redhat9 on a P-233 would be crazy? Well why would that be any more crazy than trying to install RH7 and run the software that comes with 9? That's exactly waht you're doing here. You can install RH9 (or Fedora these days) and just run a lightweight wm and you should be fine. You went about things the wrong way, assuming RH9 is different at some fundamental level from 7, which it's not.
This is the sort of thing that fascinates me. "What Linux needs to conquer the desktop" changes every year in the most astounding ways. A few years ago, it was "We need a browser as good as IE because Netscape is crap and Mozilla is slow bloated and clunky". Then it was "We need an Office replacement". Then it was "We need a good UI, because KDE and Gnome just aren't good enough".
Now it's getting so specialized. The complaints these days are "we need a real Outlook clone that's interoperable/groupware that's compatible with exchange" and that's about it. No more talk of corporate support. No more talk of good desktop UI's. No more talk of nice pretty installers. No more talk of hardware autodetection. No more talk of drivers. No more talk of good Office software, Web browsers, media players, or pretty widgets. I've got news for you, the train's been coming, slow but sure. All the old complaints are dying off and things keep moving forward. I wonder what people will say needs to happen before Linux conquers the desktop next?
What in the world are you using? LFS? Gentoo? Why are you compiling gaim? I've never compiled gaim myself and I've been a Linux and gaim user for five years now.
.exe you got from download.com than selecting a program from a list in synaptic and telling it to install doesn't mean that you can't be retrained in all of 5 minutes. I sure did that with a friend of mine and he had no troubles installing the software that came with his distro. Learn to use the tools that are there for newbies and you'll be fine. It's not really so hard, and I can personally attest to this because the learning curve has dropped significantly since I started using Linux. It's ready for those who are willing to use it.
If you sit a user down at a windows box, you'll never see them say "I want to customize the UI of this thing, give me a different window manager now!" They'll just use what's there. In the case of linux, if it's KDE2 then they'll use that. If it's KDE3 then they'll use that. If it's fvwm then they'll likely have some trouble until you show them how to work it. My largely computer illiterate friends had no troubles at all with windowmaker or icewm.
And as for dependencies, use your distro properly! Debian, Redhat, Mandrake, SuSE, ad infitum will have programs to properly manage depenancies so you don't have to. This problem was solved ages ago. apt and RPM were written well before I started using Linux, so it's not like they haven't been around out in the open for you to find.
Sure, maybe this or that distro might not have everything perfectly set up the way you want it, but then again neither does windows initially. Things still have to be installed, and just because you might be more used to double clicking on some random
Homestar Runner is one of the best sites on the net. XiaoXiaoMovie is awesome as well. I can't think of any other sites like these off the top of my head, but I'm sure others can add to the list.
And get a logo! Slashdot can't keep putting stories about you under the Redhat icon any more. Time for an identity! :-)
And best of luck with the distro too.
I personally hope Fedora actually gains some critical mass, but the problem with it seems like the same one that plagued mozilla at first. It started as a closed project, and it took a while to attract people to get that critical mass. With something like Debian or Gentoo, they grew out of openess to begin with, so by the time the project was big it already had a healthy mix of developers. Fedora will attract people, if it's really a worthwhile project, and if not then those people will move to Gentoo or Debian.
Just out of curiosity, which character in Watchmen would you be? And if you couldn't play Morpheus, who would you play in the Sandman mythos?
I'm no actor, but I'd probably go with Nightowl (the young one) from Watchmen and voicing Matthew the Raven from Sandman. The first has this great sense of being lost in the world, and I adore Matthew, especially for the way he handles things in The Wake.
It's available in experimental. It probably hasn't hit the general user populace (and really hasn't hit those outside of the Debian userbase) but experimental is slated to take on certain roles once held by unstable. Any sort of -snapshot packages (basically cvs pulls that are pre-release) now go in to experimental rather than unstable. Unstable is for packages that the maintainer considers release quality but hasn't tested widely. Experimental is for when you're still working on the thing, but would like to allow testing anyway.
At least, that's the plan. It's pretty new but it seems to be working so far. Hopefully it'll lead to faster releases.
Anyhow, the xfree 4.3 packages are in experimental. The TODO list for those packages is posted in the subversion archive where the Debian packaging development is done (you can get to it from the Debian X Strikeforce webpage) so you can see how much is left before the packages migrate to unstable. As it is though, the 4.3 packages are there if you want them.
I agree, that Debian has some problems with usability. While there are some things it does very well usability-wise, including the menu system, the alternatives system, the stock kernels, the kernel-package and patch system, and conffiles, it does lack cohesiveness in other ways. A big part of this is that the project is so damn big, it's basically impossible to wrangle anything so complex and difficult as usability guidelines (although I've considered trying to write some myself, but have no idea where to start).
On the other hand, Debian does provide options to do some things rather well. You can install discover and have hardware autodetection if you want it. You could have used PGI in the past to get yourself that graphical installer with no real fuss. The problem is the plethora of choices and sorting through them, not ignoring the users. I would very much like to see the usability of Debian extended (the new installer does have a gtk frontend, by the way) but the problem is, as usual, both related to politics and manpower.
Ultimately, I'd love to hear suggestions on ways that Debian could improve usability. I personally think it's a power user's dream come true in terms of usability. How would you improve it, beyond a graphical installer? Hardware autodetection is also, I believe, a standard component of the new installer, so don't list that please.
Not really. The mplayer situation has been progressing, especially because some of the flames have calmed down, and the mplayer developers have been a bit nicer on the Debian mailing lists. There are still major hurdles to clear, mainly social ones. Many Debian developers, including those that administer the ftp archive, don't trust the mplayer developers yet. Hopefully it'll get in at some point though.
Yes, and it's about time too. I'm a little hesitant about the international users issue that kept coming up during the last discussion about removing non-free, but compared to the days of yore when we couldn't contemplate life without netscape, this is a small thing, and will probably fix itself when the need arises. I can almost see it as the project coming to fruition in a way.
And then once the software is in the hands of the distro makers, they have to package it properly. Now, believe it or not, but packaging something well is tough. It's like programming, so you have to shake the bugs out of that too over time.
I guess it comes down to a different definition of stable. If you think all software is stable right out of the gate because it runs, then you've got a different definition of stable than most Debian developers. If you want that stuff, there's a version of Debian you can track: it's called unstable (and you can even sprinkle in experimental, if you really have the faith).
I really don't understand the general Linux user's need to get the latest and greatest at all times. Most of us ran windows for years, and simply waited for Microsoft to say "Ok, it's ready, here you go" every two or more years before upgrading. Debian's release schedule isn't a whole lot different, but you can simply see what's going on behind the scenes, so people tend to get impatient. How quickly we forget.
Because it goes against the very philosophy of Debian, right down to its initial charter if you want to go read it. Branden is one of the few people who actually still give a damn about what the project was originally founded for. He produces quality work (the Debian X packages are top notch, and he basically manages the porting work to other arch's because XFree puts out unportable crap) and he sticks to his ideals in the face of criticism.
He's not stopping you from using qmail, nor is he stopping people from putting up their own apt sources (people do it every day!) that host non-free, he just wants it out of the project so it can be free of this bizarre dichotomy.
No. No, not at all. In fact it moves them farther away from Gentoo. In Gentoo's Social Contract, there is nothing explicitly stated about Documentation, but rather refers explicitly to software as binaries or sources. Debian has been working on productive discussions with the FSF over the GFDL for over two years, and this change is a direct result from those discussions. Most Debian Developers feel that documentation qualifies as software, and should be included under the DFSG as well.
Everything in Gentoo's Social contract is basically directly lifted from Debian's, although they decided not to take it all. The Gentoo people don't operate on nearly the same strict standards of Freedom that Debian does, and the differences in the Social Contracts, including the latest change, demonstrates that. If the Gentoo people decide to move in this direction too, it'll be because of more than two years of hard wrangling on debian-legal.
Historical reasons, apparently. When exim was chosen as the default MTA for Debian, postfix's license was in review and it wasn't sure as to whether or not it satisfied the DFSG. There was a discussion about changing the default MTA for the next release (here is the thread) but no real progress was made. exim4 seems like it'll be a perfectly fine choice, especially since no truly compelling argument could be made to switch, as much as I'd root for postfix myself.
Because they're not really winning their way in to the curriculum based on merit, but on donations. Does that seem right to you?
Any serious CS student will find out about Linux anyway, and any serious CS student will try it out. If they like it, they'll keep using it, and if not, they'll go back to Windows. If they like it enough to contribute, then they'll learn to program it, and if they don't then they'll just program what they learned in class.
Now how is this any different than the way things already were?
The thing about apt, at least for Debian users, is that there is a strong and established mirror system already in place that handles the load just fine. I don't know how this applies to aptrpm users and whatnot.
As a sidenote, one of the main dpkg developers was playing around with using bittorrent in conjunction with the mirror system for a while, so the problem has at least been toyed with. I don't know if we'll ever see it in action though, since the need just isn't there right now.
That's only because sharecropping is considered the standard right now. It doesn't have to be that way forever, nor should it be.
Just because mass market PC products are made for windows doesn't mean that they won't run Linux just fine. You'd have a very difficult time demonstrating that Windows can run on more hardware configurations than Linux, especially given that Linux supports totally different architectures in addition to x86.
Not to speak for the parent poster, but Skylarov would probably be the big issue that anyone would hold against Adobe, and it is a rather large issue.