Actually, this project has been in the works for years, well before Redhat even announced that they were ending their free distro.
The primary motivation for the new installer was that the old one had a horrific codebase that no one wanted to touch. It was a major sticking point for the last release as to whether or not the focus should be on the new installer (then very much in its infancy) or "just" polishing up the old one and shoving it out the door. The latter choice was made, and it turned out not to be worthwhile.
This new installer is much nicer under the hood, in that it's made up of individual components that can be swapped in and out relatively easily. Once recent example of this is a few months ago the installer switched from the old partconf partitioner to the newer partman that you see in it now. This was a very easy and smooth transition, thanks to the way the new installer is structured.
The other advantage to the new installer was that it was a good chance to implement things that the users were asking for, including hardware autodetection, aptitude instead of dselect, grub over lilo, wifi autodetection, less questions, etc. There's still lots of requests that have to be filled in. pppoe support is only in its infancy, 2.6 needs a lot more testing, the documentation needs a lot of work, and some multilanguage issues need to be solved before a gtk interface can be slapped on. Those are only a few of the holes that I personally see, I'm sure that other people on the team can bring up others. Ultimately though, I know the core members of this team and I can definitely say that they're not concerned with Redhat at all. They, and I, simply want to build the best installer possible for Debian so that we can not only release sarge soon, but also prevent the installer from being a reason for holding up future releases. We're crafting this one for our own future.
Because 2.4 has been heavily tested within the installer, so we know it's good. 2.6 has only just been put in, and it needs a whole lot more testing before it makes sense as the default. We'll still provide it, of course, as an alternative boot option, but there's no reason to break the installer just so the shiny new toy can be the default when the old one works perfectly well.
Yes, the new installer will autodetect and select the right values for your X config. We've had a lot of successful reports about it so far, although more testing would definitely be a good thing.
I want to be able to download a program, click it and it be on my menu and ready to use.
Linux has worked this way for years, and now basically all distros really do work this way. Instead of a model where you go to some website and download a program that way, you use a different program that actually grabs all the dependencies for you. Which one you use depends on your distro, but for newbie distros you'll want yum for Fedora, urpmi for Mandrake, and synaptic for Debian derivatives like Knoppix and Mepis (although aptitude is a fantastic choice for these as well). They'll give you a list of available programs and you just tell it what you want to install. Simple as that, and you have the advantage of a specialized app that takes care of the whole process for you rather than having to do it manually like in Windows.
Seriously, installing apps in Linux is actually easier than Windows. It just doesn't behave like Windows, so people get frustrated because they foolishly try and do all the work themselves.
The model you use in linux is different in that your distro provides a whole library of programs for you to download and you go through them. If you do, it should work flawlessly the way you describe. It's just a shift in mindset from the download.com way of doing things.
I actually thought this distinction was the shining point in the article, in that it actually contributed something new to the discussions about desktop linux that have been going on for ages. It's not about locking down the box so much as needs of users. When you hear the debates, you hear the two sides saying "Linux now has a good office suite, email client, etc" while the other side says "yeah, but advanced Office users need their Excel macros and their Outlook calendars".
To me, this difference was basically given terminology by this article. The people who need their Excel Macros and aren't ready to switch over are the Information Workers while the ones who just need to type a few emails and memos are the Transaction Workers. It basically clarifies the fact that some people will do just fine with a Linux desktop while others aren't ready. We all know this, but no one's given names to define this distinction before.
To me, it's incredible to see this distinction finally being raised because 5 years ago you couldn't really say that Linux was ready for either. Progress is happening.
Because not everybody wants to use Debian or Gentoo.
No one said they should have to.
Because the model in which huge numbers of people try and package every single thing in a centralised manner is inefficient and doesn't scale. How many levels does Debian have now? 4 - stable, testing, unstable and experimental, iirc.
The model has scaled pretty well, as far as I can see. Not every piece of crap program on freshmeat needs to be available for the novice user, nor should it be. The model might not be the most efficient one possible, but it definitely produces the most robust system, which can not be said for something like Redhat's contrib packages.
The different levels of Debian are due to more complex issues than the large packaging model, including high standards of quality and porting issues. It's also an ongoing experiment in massive collaboration, and the jury is still out on whether testing or experimental as they're being used are worthwhile.
The fact is, that it takes a lot of time to package everything like that, and a lot of human resources, which most distros simply cannot afford.
Indeed, but from what I hear, Fedora and Gentoo have solved these issues as well. Manpower is available in abundance thanks to the FOSS development model, that's not the limiting factor.
Even there, in order to get up to date software you must move on to progressively more and more unstable sets of packages, which really is just not acceptable for most people.
Debian's model offers you a choice, and this is one you simply can't escape. If you want the new stuff, be prepared to put up with a few bugs. If you want the stability, be prepared to put up with old software. It's up to you. I've used unstable exclusively on my desktop for many years with very few issues. The software itself is as stable as anything packaged and put out by other distros at the same time points. Debian stable is a level of quality and reliability that other distros simply don't offer (now RHEL does, but that's in a different category). If you run Fedora and you're worried about stability, you may as well be running unstable. Debian simply offers you a choice.
And as for it being a solved problem, it is in my, as well as many others' experience. If you choose to run your machine in such a way that it hampers your work and frustrates you, be my guest. The choice is yours, but don't complain that it's so difficult when you chose that path.
This bothered me too, but maybe that's because I've been using Debian for so long. Seriously, do Linux users really work like this? Hunting after the right RPM's and such? I mean... I remember doing that in 1999 (and it was a solved problem even then) on Mandrake, but they had already begun work on urpmi or something similar at the time. Honestly, I haven't dealt with a real dependency issue in half a decade and I haven't had to compile an app that I didn't write in years (unless I wanted to package it myself for some reason). I thought this was long solved for all distros now.
Don't forget that *ONE* scientist had a dream about the structure of DNA
You mean Watson and Crick? How about Rosalin Franklin? How about Maurice Wilkins? Not to mention all the other people who indirectly helped define the structure including Linus Pauling, Owen Chargaff, and Oswald Avery? You're forgetting a fairly large cast, and I recommend reading The Double Helix to get some basics down about this particular story.
I don't know as much about Einstein, but he didn't unify physics, and I'm sure that the field couldn't have gotten far without the likes of Neils Bohr and the many others who helped define quantum mechanics.
Innovation never happens only as the result of one lone genius pushing a field forward. Google for "standing on the shoulders of giants" for a good quote on the topic.
Well, the major Debian-specific discussions about the new X license were in reference to the X-Oz license, which is very similar. The representative for X-Oz was very unclear about the questions that the Debian Developers had as to how the license applies, so the current decision that the X-Oz license, and by proxy the XFree86 1.1 license are not DFSG compliant would have to stand. Note that the representative said she would get back to the list after a week. This was about a month ago. As a result, the packages would likely have to go in to non-free unless some sort of clarification on the license could be achieved.
This is the basic ruling on the X-Oz license as it stands now, and a following email in the thread lists the differences between X-Oz and the XFree86 1.1 licenses.
Yes. I personally use the kernel-image-2.6-686 package, which grabs the newest kernel once it's available automatically. It updates lilo.conf for me and everything. I rolled my own kernels for about 4 years before switching to the stock Debian ones, and I have to admit that the Debian kernel maintainer (Herbert Xu) does a great job. Things always work and I don't have to hassle with my own mistakes. It also frees up my time to work on other things, which is what Debian is supposed to be about.
The Debian install did NOT configure my Soundblaster Live soundcard, pissing me off quite a bit.
The new installer for sarge, currently in beta3, will detect and set up your sound card just fine. The installer has been rewritten from scrath and is much improved. Trust me, we've all been aware of how much the old installer sucked, which is why it's been rewritten.
I decided if the much heralded dist-upgrade command sequence worked to upgrade to unstable I would become a Debian zealot and forgive the soundcard hassle. It upgraded to KDE 3.2.something, but borked my computer.
That could be a for any number of reasons. One is that unstable is unstable, and development versions don't always work as expected on that particular day. You probably would have been better off with a backport. Also, for future reference, apt-get isn't really equipped to handle a major upgrade for a couple of reasons. Much of the propaganda surrounding it is far overrated. You're much better off with a more functional apt frontend like aptitude or synaptic.
Finally, apt isn't all it's cracked up to be. It's just a package management system. The real benefit of Debian is in the quality of the packages themselves. urpmi is no different in this regard. You went from one stable version of Mandrake to another stable version of Mandrake with no hassles. Apt can do the same thing, but you went from a stable version of Debian to a development version and caught unstable on what was probably a bad day. The real issue in this regard is slow releases, but backports are there to help with that. I'm glad you're happy with Mandrake though (it's a great distro) and hopefully you'll try Debian again one day.
Please send an email to debian-release or debian-project about this. I think having torrents set up would be a great idea, but the developers need to hear it.
The software was really out of date even with the so called "unstable" version.
Uh... what software was this, exactly? XFree was the lone exception for unstable for a long time, but that's essentially up to date now (unless you count 4.4, which no one but slack iirc is shipping).
Then I went on irc to see the community and between the "rms is a blowhard" and "debian unstable is way more stabler than redcrap, teee ehehe" people
Well, in terms of the former, RMS has pissed off the Debian community quite a bit lately. Between the GFDL fiasco and his labeling Debian as not free enough because it has the non-free section of the archive, he's not been too kind to what is undoubtedly the distro most concerned with Free Software as such.
As for the latter, there's all sorts of distro bashing in any forum. That's the way it is. It's called friendly rivalry. If you actually look at what's going on above the IRC level, there's a lot of real cooperation going on between the distros, for all the petty rivalry. Lots of Debian Developers, for instance, are employed by Redhat.
there was their ringleader one guy some mwilson that would just make fun of anyone who asked a question calling them "morons, cluetards, braindead," etc. and since he was flaming about 10 people at once like some kind of burning octopus of negativity I know I am not the only person experiencing that.
Yes, mwilson is a complete and utter ass hole. Yes, he's known to be as such. But you're judging a whole channel based on one guy. You do have/ignore, as well as the power of your own brain, to ignore people like that. I recommend them both.
The debian community apparently lets the trolls run it's irc channel.
The Debian Developers have, as a whole, written off #debian. I think most developers would want to see it as totally separate from the project as a whole, which at this point it probably is. A major reason for that is that the users don't let the project know that they want a good IRC channel where they can get help. Most developers see it as useless. If you want the channel to be more tightly regulated by the project, I recommend sending a mail to debian-project and letting them know how you feel. If there's enough people who really want the channel to be policed differently and brought more in to the fold of the project itself, please speak up so you can actually influence things. Unless you'd rather just complain on slashdot more.
So from my experience anyways it isn't really a distro to be taken seriously. If want to use old software and get flamed by some ultra-leet dudes on irc then go ahead use debian. If you want something else a tad more professional use basically anything else.
Way to judge an entire distribution based on its IRC channel. Talk about professional!
create a friendly user community that doesn't slam people for asking questions "improperly"
This is a difficult thing. I hang around in #debian quite a bit, and I see both the types of people who are happy to help and those who are only there to make themselves feel smarter and better than the rest. And I can honestly say that the former are more plentiful than the latter, but users tend to actually listen to the dicks. Why? I don't know... maybe they really would rather get in a pointless fight than get something fixed, but it's something I see all the time. The trick is to listen to those who are trying to help you, use/ignore when necessary, and simply grow a thicker skin.
On the other side, as someone who tries to help out with support, I see two general kinds of users. One is the type who just wants their hand held all the way through the process. They'll not bother to do a couple of google searches or search the Debian list archives. They won't bother with the manpage because "it takes too long to read" (as though logging in to IRC, asking your question, clarifying your question, and waiting for a response is any faster). Sometimes they'll even lie about what they actually did to their system. These people are amazing and get ridiculed. Those who troll and say "Debian sucks! I can't get this piece of crap to work, so I'm going to gentoo!" (as many people on slashdot recommend) tend to be astounded when the channel unanimously says "Ok, see ya later!" These people aren't treated very well, but that's because they're not treating anyone else very well in return.
On the other hand, users who ask smart questions and are willing to do some basic searches and reading are helped quite a bit. Many people in #debian genuinely try to help these people, because we can all see ourselves in this position. You're not guaranteed an answer, or even help (oftentimes no one can help with your question) but if you're willing to do a little bit of your own free tech support then you'll do just fine. Remember, when someone points you to a manpage or the Installation Manual (yes, there is one, and it's amazing how few people read it) it's not because they're being a dick, it's because the doc really does have the answer. An "RTFM" always comes with a pointer towards what to read in my experience, and if you're willing to listen you'll do fine.
If someone thinks a question is unworthy they should not waste their time by finishing reading it and they certainly shouldn't spend their time answering the question. Doing and complaining, about either given the voluntary nature of the internet makes them look like a mean loser.
Agreed. How do you expect to change that though? Debian is, by nature, a very open project, and as such places like debian-user and #debian are relatively unmoderated. Do you want to close off the lists somehow? Do you want to moderate them? Who do you propose to do the job? And by what guidelines? And by what mechanism? These are serious questions, and they don't have easy answers. "Fixing" a community isn't easy, and if you want to help I suggest hopping on debian-user or stopping by #debian and try and change things one user at a time.
Could you be more specific? How could the install be simpler? If you just use the installer in basic ways (like using pre-supplied partition schemes and tasksel) you can get a full Debian install pretty easily at present.
I download PHP and some other tools to get a web site running, wham, something doesn't work, research, research, finally find some note that one author made a change to one module that breaks PHP support, but the PHP folks say it will be fixed in the next version.
You're making a site and all of a sudden it stops working while you're in the middle of developing it? What'd you do? Upgrade one of those tools? Just downgrade it until the fix you need is released. Geez, this is basic admininstration here, whether the source is open or closed.
When people bitch about how long Debian takes to release a new version, now you know exactly why. It's hard to get software in really solid shape in and of itself, and then on top of that you have to get the packages working together nicely. This is hard work, and you've now seen why. Sure, you can always grab pre-release packages from outside sources, but these haven't gone through the Q&A that your distro provides. Packages coming from within the distro itself should play nicely together. That's the point of a distro, after all.
If you want to suggest these things to your bosses, be prepared to live with the tradeoffs. You can have stable software that's nice, but you'll pay the price in that it won't be shiny and new. Or you can have the new stuff, but be prepared to play "perpetual beta tester".
Any Linux install is easy to fuck up, if you try hard enough. You obviously tried very hard to fuck yours up, and did a good job of it. If you're suggesting Linux to a professional admin, hopefully they'll be a little more clue'd in about how their system works than you are, and will be able to deal with their system properly. As an example, I run Debian unstable on my desktop, a system which is known for having bugs pop up from time to time. But I know how to deal with them and it's never ever amounted to a reinstall, and only about twice in the last four years has it even interrupted my workflow. As another example, Gentoo allows you to completely fuck up your machine if you want very easily, and yet tons of people can't stop gushing about how easy it is to use. If you know how to deal with potential problems, Linux is an amazing choice, mainly because those problems are relatively transparent compared to something like Windows.
Oh, and I don't know how it is for Fedora or Mandrake, but in Debian, the majority of the docs that you'll ever want are located in/usr/share/doc/packagename or the program's manpage (if a program doesn't have a manpage, that's considered a bug). It's very rare that I have to go outside those two sources to figure out how to do something on my system.
We're using some standard tools like hotplug, as well as stuff that is more or less standard in Debian but not standard outside of it like Progeny's discover for the autodetection. Discover is a mature program, and is working rather well, and we'll see if the newest version of it makes it in to the installer before release.
The installer also uses a few other apps, like mdetect and read-edid (x86 only) to help configure the XFree86 packages, although the X configuration stuff is going to be rewritten soon, and the initial work has started on that. Nothing has been actively taken from knoppix that I'm aware of though.
Wow, that article brings back memories. Anyhow, the fact that the population doesn't use or know about mozilla or firefox right now isn't a big deal. They'll continue to use IE because that's what they know. They even do this on the Mac (which is really sick, given IE's crappy state on the Mac).
When Linux starts to move in to more and more corporate desktops, people won't be able to rely on their IE habit anymore, and will be forced to use a Linux browser. This is when you'll start to see it in greater force.
And even further down the line, when Linux starts to invade the home desktop space also, we'll be glad the Mozilla project (and the KDE/KHTML project) has been around for so long. These things will come, it's only a matter of time.
How many distros even have a company sponsored community list serv?
Debian does. Has for a long time. It has a lot of company sponsored servers for just about everything, as well as mirrors. Plus there are university sponsored ones, and a few private machines out there as well for development tasks.
Sure it is, with people like you all whining about a little compile time, less new people will try it. So shut it.
Wow, that makes me trust your opinion. The compile time really is an issue with gentoo. If you don't mind it, then it's fine, but if you do then you should be aware of it. There's no One True Distro, thankfully, nor will there ever be.
Dude, did someone take away your copy of fvwm or something? Mutt somehow just grew a GUI and started autoexecuting content? Are you now bored with LyX and want to go back to Microsoft Office? Does apt-get now only work through your web browser and only download certain pieces of software? Hell, even Gnome and KDE are doing interesting stuff these days if you've been paying attention. They've gone way past the "Windows Clone" stage, if they ever really were in one.
Linux is still cool, and it gives you plenty of stuff to play with that's still a pain in the ass to get working in Windows (and even then, it's only going through cygwin). No one stopped the programs that were so different then from being different now. twm still works great here.
I have a hard time recommending testing to anyone. At the very least, you have to subscribe to the debian security announcements list, just to be sure you're Ok (which is something you should be doing no matter which version of Debian you run). But my experiences in unstable echo Bruce's. I've used it continually for four years straight, and I've had only one major break, and that was over three years ago. If you're sufficiently clued in to things (like knowing how to check the BTS) unstable is amazingly solid, especially for a development branch.
You ought to send a bug report to the discover maintainers about the firewire thing. They're very quick to respond to them these days. discover only handles cards though, not hotplugged devices like mice and USB stickes, that's what hotplug is for. If your firewire card isn't working properly, use reportbug to send a report to discover-data, with the correct module to load, as well as the output from lspci and lspci -n.
Oh, and the 2.6 images have ACPI compiled in, although the discover support for 2.6 is kinda sketchy at present.
Slashdot Mirror
Actually, this project has been in the works for years, well before Redhat even announced that they were ending their free distro.
The primary motivation for the new installer was that the old one had a horrific codebase that no one wanted to touch. It was a major sticking point for the last release as to whether or not the focus should be on the new installer (then very much in its infancy) or "just" polishing up the old one and shoving it out the door. The latter choice was made, and it turned out not to be worthwhile.
This new installer is much nicer under the hood, in that it's made up of individual components that can be swapped in and out relatively easily. Once recent example of this is a few months ago the installer switched from the old partconf partitioner to the newer partman that you see in it now. This was a very easy and smooth transition, thanks to the way the new installer is structured.
The other advantage to the new installer was that it was a good chance to implement things that the users were asking for, including hardware autodetection, aptitude instead of dselect, grub over lilo, wifi autodetection, less questions, etc. There's still lots of requests that have to be filled in. pppoe support is only in its infancy, 2.6 needs a lot more testing, the documentation needs a lot of work, and some multilanguage issues need to be solved before a gtk interface can be slapped on. Those are only a few of the holes that I personally see, I'm sure that other people on the team can bring up others. Ultimately though, I know the core members of this team and I can definitely say that they're not concerned with Redhat at all. They, and I, simply want to build the best installer possible for Debian so that we can not only release sarge soon, but also prevent the installer from being a reason for holding up future releases. We're crafting this one for our own future.
Because 2.4 has been heavily tested within the installer, so we know it's good. 2.6 has only just been put in, and it needs a whole lot more testing before it makes sense as the default. We'll still provide it, of course, as an alternative boot option, but there's no reason to break the installer just so the shiny new toy can be the default when the old one works perfectly well.
Yes, the new installer will autodetect and select the right values for your X config. We've had a lot of successful reports about it so far, although more testing would definitely be a good thing.
Seriously, installing apps in Linux is actually easier than Windows. It just doesn't behave like Windows, so people get frustrated because they foolishly try and do all the work themselves.
The model you use in linux is different in that your distro provides a whole library of programs for you to download and you go through them. If you do, it should work flawlessly the way you describe. It's just a shift in mindset from the download.com way of doing things.
I actually thought this distinction was the shining point in the article, in that it actually contributed something new to the discussions about desktop linux that have been going on for ages. It's not about locking down the box so much as needs of users. When you hear the debates, you hear the two sides saying "Linux now has a good office suite, email client, etc" while the other side says "yeah, but advanced Office users need their Excel macros and their Outlook calendars".
To me, this difference was basically given terminology by this article. The people who need their Excel Macros and aren't ready to switch over are the Information Workers while the ones who just need to type a few emails and memos are the Transaction Workers. It basically clarifies the fact that some people will do just fine with a Linux desktop while others aren't ready. We all know this, but no one's given names to define this distinction before.
To me, it's incredible to see this distinction finally being raised because 5 years ago you couldn't really say that Linux was ready for either. Progress is happening.
The model has scaled pretty well, as far as I can see. Not every piece of crap program on freshmeat needs to be available for the novice user, nor should it be. The model might not be the most efficient one possible, but it definitely produces the most robust system, which can not be said for something like Redhat's contrib packages. The different levels of Debian are due to more complex issues than the large packaging model, including high standards of quality and porting issues. It's also an ongoing experiment in massive collaboration, and the jury is still out on whether testing or experimental as they're being used are worthwhile.
Indeed, but from what I hear, Fedora and Gentoo have solved these issues as well. Manpower is available in abundance thanks to the FOSS development model, that's not the limiting factor.
Debian's model offers you a choice, and this is one you simply can't escape. If you want the new stuff, be prepared to put up with a few bugs. If you want the stability, be prepared to put up with old software. It's up to you. I've used unstable exclusively on my desktop for many years with very few issues. The software itself is as stable as anything packaged and put out by other distros at the same time points. Debian stable is a level of quality and reliability that other distros simply don't offer (now RHEL does, but that's in a different category). If you run Fedora and you're worried about stability, you may as well be running unstable. Debian simply offers you a choice.
And as for it being a solved problem, it is in my, as well as many others' experience. If you choose to run your machine in such a way that it hampers your work and frustrates you, be my guest. The choice is yours, but don't complain that it's so difficult when you chose that path.
This bothered me too, but maybe that's because I've been using Debian for so long. Seriously, do Linux users really work like this? Hunting after the right RPM's and such? I mean... I remember doing that in 1999 (and it was a solved problem even then) on Mandrake, but they had already begun work on urpmi or something similar at the time. Honestly, I haven't dealt with a real dependency issue in half a decade and I haven't had to compile an app that I didn't write in years (unless I wanted to package it myself for some reason). I thought this was long solved for all distros now.
Do you guys really live like this? If so... why?
I don't know as much about Einstein, but he didn't unify physics, and I'm sure that the field couldn't have gotten far without the likes of Neils Bohr and the many others who helped define quantum mechanics.
Innovation never happens only as the result of one lone genius pushing a field forward. Google for "standing on the shoulders of giants" for a good quote on the topic.
Well, the major Debian-specific discussions about the new X license were in reference to the X-Oz license, which is very similar. The representative for X-Oz was very unclear about the questions that the Debian Developers had as to how the license applies, so the current decision that the X-Oz license, and by proxy the XFree86 1.1 license are not DFSG compliant would have to stand. Note that the representative said she would get back to the list after a week. This was about a month ago. As a result, the packages would likely have to go in to non-free unless some sort of clarification on the license could be achieved.
This is the basic ruling on the X-Oz license as it stands now, and a following email in the thread lists the differences between X-Oz and the XFree86 1.1 licenses.
Yes. I personally use the kernel-image-2.6-686 package, which grabs the newest kernel once it's available automatically. It updates lilo.conf for me and everything. I rolled my own kernels for about 4 years before switching to the stock Debian ones, and I have to admit that the Debian kernel maintainer (Herbert Xu) does a great job. Things always work and I don't have to hassle with my own mistakes. It also frees up my time to work on other things, which is what Debian is supposed to be about.
That could be a for any number of reasons. One is that unstable is unstable, and development versions don't always work as expected on that particular day. You probably would have been better off with a backport. Also, for future reference, apt-get isn't really equipped to handle a major upgrade for a couple of reasons. Much of the propaganda surrounding it is far overrated. You're much better off with a more functional apt frontend like aptitude or synaptic.
Finally, apt isn't all it's cracked up to be. It's just a package management system. The real benefit of Debian is in the quality of the packages themselves. urpmi is no different in this regard. You went from one stable version of Mandrake to another stable version of Mandrake with no hassles. Apt can do the same thing, but you went from a stable version of Debian to a development version and caught unstable on what was probably a bad day. The real issue in this regard is slow releases, but backports are there to help with that. I'm glad you're happy with Mandrake though (it's a great distro) and hopefully you'll try Debian again one day.
Or the more user-friendly configure-debian?
Well, in terms of the former, RMS has pissed off the Debian community quite a bit lately. Between the GFDL fiasco and his labeling Debian as not free enough because it has the non-free section of the archive, he's not been too kind to what is undoubtedly the distro most concerned with Free Software as such.
As for the latter, there's all sorts of distro bashing in any forum. That's the way it is. It's called friendly rivalry. If you actually look at what's going on above the IRC level, there's a lot of real cooperation going on between the distros, for all the petty rivalry. Lots of Debian Developers, for instance, are employed by Redhat.
Yes, mwilson is a complete and utter ass hole. Yes, he's known to be as such. But you're judging a whole channel based on one guy. You do have
The Debian Developers have, as a whole, written off #debian. I think most developers would want to see it as totally separate from the project as a whole, which at this point it probably is. A major reason for that is that the users don't let the project know that they want a good IRC channel where they can get help. Most developers see it as useless. If you want the channel to be more tightly regulated by the project, I recommend sending a mail to debian-project and letting them know how you feel. If there's enough people who really want the channel to be policed differently and brought more in to the fold of the project itself, please speak up so you can actually influence things. Unless you'd rather just complain on slashdot more.
Way to judge an entire distribution based on its IRC channel. Talk about professional!
On the other side, as someone who tries to help out with support, I see two general kinds of users. One is the type who just wants their hand held all the way through the process. They'll not bother to do a couple of google searches or search the Debian list archives. They won't bother with the manpage because "it takes too long to read" (as though logging in to IRC, asking your question, clarifying your question, and waiting for a response is any faster). Sometimes they'll even lie about what they actually did to their system. These people are amazing and get ridiculed. Those who troll and say "Debian sucks! I can't get this piece of crap to work, so I'm going to gentoo!" (as many people on slashdot recommend) tend to be astounded when the channel unanimously says "Ok, see ya later!" These people aren't treated very well, but that's because they're not treating anyone else very well in return.
On the other hand, users who ask smart questions and are willing to do some basic searches and reading are helped quite a bit. Many people in #debian genuinely try to help these people, because we can all see ourselves in this position. You're not guaranteed an answer, or even help (oftentimes no one can help with your question) but if you're willing to do a little bit of your own free tech support then you'll do just fine. Remember, when someone points you to a manpage or the Installation Manual (yes, there is one, and it's amazing how few people read it) it's not because they're being a dick, it's because the doc really does have the answer. An "RTFM" always comes with a pointer towards what to read in my experience, and if you're willing to listen you'll do fine.
Agreed. How do you expect to change that though? Debian is, by nature, a very open project, and as such places like debian-user and #debian are relatively unmoderated. Do you want to close off the lists somehow? Do you want to moderate them? Who do you propose to do the job? And by what guidelines? And by what mechanism? These are serious questions, and they don't have easy answers. "Fixing" a community isn't easy, and if you want to help I suggest hopping on debian-user or stopping by #debian and try and change things one user at a time.
Could you be more specific? How could the install be simpler? If you just use the installer in basic ways (like using pre-supplied partition schemes and tasksel) you can get a full Debian install pretty easily at present.
When people bitch about how long Debian takes to release a new version, now you know exactly why. It's hard to get software in really solid shape in and of itself, and then on top of that you have to get the packages working together nicely. This is hard work, and you've now seen why. Sure, you can always grab pre-release packages from outside sources, but these haven't gone through the Q&A that your distro provides. Packages coming from within the distro itself should play nicely together. That's the point of a distro, after all.
/usr/share/doc/packagename or the program's manpage (if a program doesn't have a manpage, that's considered a bug). It's very rare that I have to go outside those two sources to figure out how to do something on my system.
If you want to suggest these things to your bosses, be prepared to live with the tradeoffs. You can have stable software that's nice, but you'll pay the price in that it won't be shiny and new. Or you can have the new stuff, but be prepared to play "perpetual beta tester".
Any Linux install is easy to fuck up, if you try hard enough. You obviously tried very hard to fuck yours up, and did a good job of it. If you're suggesting Linux to a professional admin, hopefully they'll be a little more clue'd in about how their system works than you are, and will be able to deal with their system properly. As an example, I run Debian unstable on my desktop, a system which is known for having bugs pop up from time to time. But I know how to deal with them and it's never ever amounted to a reinstall, and only about twice in the last four years has it even interrupted my workflow. As another example, Gentoo allows you to completely fuck up your machine if you want very easily, and yet tons of people can't stop gushing about how easy it is to use. If you know how to deal with potential problems, Linux is an amazing choice, mainly because those problems are relatively transparent compared to something like Windows.
Oh, and I don't know how it is for Fedora or Mandrake, but in Debian, the majority of the docs that you'll ever want are located in
If that really is an accurate figure, then things are really improving. I, for one, hope so.
We're using some standard tools like hotplug, as well as stuff that is more or less standard in Debian but not standard outside of it like Progeny's discover for the autodetection. Discover is a mature program, and is working rather well, and we'll see if the newest version of it makes it in to the installer before release.
The installer also uses a few other apps, like mdetect and read-edid (x86 only) to help configure the XFree86 packages, although the X configuration stuff is going to be rewritten soon, and the initial work has started on that. Nothing has been actively taken from knoppix that I'm aware of though.
Wow, that article brings back memories. Anyhow, the fact that the population doesn't use or know about mozilla or firefox right now isn't a big deal. They'll continue to use IE because that's what they know. They even do this on the Mac (which is really sick, given IE's crappy state on the Mac).
When Linux starts to move in to more and more corporate desktops, people won't be able to rely on their IE habit anymore, and will be forced to use a Linux browser. This is when you'll start to see it in greater force.
And even further down the line, when Linux starts to invade the home desktop space also, we'll be glad the Mozilla project (and the KDE/KHTML project) has been around for so long. These things will come, it's only a matter of time.
Dude, did someone take away your copy of fvwm or something? Mutt somehow just grew a GUI and started autoexecuting content? Are you now bored with LyX and want to go back to Microsoft Office? Does apt-get now only work through your web browser and only download certain pieces of software? Hell, even Gnome and KDE are doing interesting stuff these days if you've been paying attention. They've gone way past the "Windows Clone" stage, if they ever really were in one.
Linux is still cool, and it gives you plenty of stuff to play with that's still a pain in the ass to get working in Windows (and even then, it's only going through cygwin). No one stopped the programs that were so different then from being different now. twm still works great here.
I have a hard time recommending testing to anyone. At the very least, you have to subscribe to the debian security announcements list, just to be sure you're Ok (which is something you should be doing no matter which version of Debian you run). But my experiences in unstable echo Bruce's. I've used it continually for four years straight, and I've had only one major break, and that was over three years ago. If you're sufficiently clued in to things (like knowing how to check the BTS) unstable is amazingly solid, especially for a development branch.
You ought to send a bug report to the discover maintainers about the firewire thing. They're very quick to respond to them these days. discover only handles cards though, not hotplugged devices like mice and USB stickes, that's what hotplug is for. If your firewire card isn't working properly, use reportbug to send a report to discover-data, with the correct module to load, as well as the output from lspci and lspci -n.
Oh, and the 2.6 images have ACPI compiled in, although the discover support for 2.6 is kinda sketchy at present.