Debian Desktop Subproject Launched 387
MrOutlander writes "The Debian Project is now officially addressing its usability on the desktop with the launch of the Debian Desktop subproject. Great to see usability being recognized as a very important part of debian. Other than the sometimes daunting install process, Debian is one of the best linux distributions."
one of the best linux distributions? (Score:5, Funny)
In other news:
-Lemonade is one of the best beverages
-Pink is the best color
-Pi is the best number
Required Reading... (Score:5, Informative)
Re:Required Reading... (Score:4, Interesting)
As someone who's used OS8, OS9, Linux, and all the video variants, let me tell you that OS8 comes pretty damn close to being *golden*. Apple spent a lot of time making OS7-9 pleasant and easy to use and it shows.
Re:Required Reading... (Score:2)
Re:Required Reading... (Score:2, Insightful)
Re:Required Reading... (Score:4, Insightful)
Yes. Yes. Yes. People..read this line and chant it.
And in the process, they could also team up with the 'Debian Graphical Installer' group (see this [slashdot.org]
Re:Required Reading... (Score:4, Interesting)
Re:Required Reading... (Score:5, Funny)
I'm not interested in interfacing with a fabled race of dwarflike creatures, I'm interested in interfacing through a GUI to a UNIX or UNIX-like system.
Re:Required Reading... (Score:2)
Oh BTW, GNOME is a desktop environment. Now you know.
Re:Required Reading... (Score:3, Funny)
Re:Required Reading... (Score:2)
Re:Required Reading... (Score:2)
Well, I do have some missing underpants. [aol.com]
Re:Required Reading... (Score:2, Funny)
Okay, you've opened it up now. Someone mentioned Gnome, you mentioned underpants, so you asked for it.
I feel dirty after posting that.
Re:Required Reading... (Score:2, Flamebait)
1. Create Debian desktop distribution
2. ???
3. Microsoft goes bankrupt!
Re:Required Reading... (Score:2)
Ok, seriously, Gnome, KDE et al have some serious usability issues to deal with. While at first they seem pretty much like Windows (or whatever else they're trying to duplicate), when you start to use them every day you begin to see how disconnected they are from the rest of the OS. No Linux interface that exists today provides unified system usability.
Re:Required Reading... (Score:3, Insightful)
Re:Required Reading... (Score:5, Insightful)
Or for that matter, windows. KDE has its faults, to be sure, but Konq provides better integration than even Windows. In Windows, I can only see folders on the left pane in explorer. In Konquerer, I can at my option see files, and view all kinds of content, from HTML to text to postscript, in the right hand pane. I can rubberband a bunch of files in ftp and drag them to the desktop. The control center includes system management (to some degree), whereas in windows, you use a completely separate app, MMC. Granted, Windows is carrying along legacy cruft, and would probably make every control panel a MMC snap-in nowadays, but they don't even provide an adaptor.
The user experience on windows is pretty disjointed too. But I don't think Unix (it's more than Linux, folks) exactly has further to go than Windows. It's just broken in different places.
Re:Required Reading... (Score:5, Funny)
Blasphemy!
Re:Required Reading... (Score:4, Insightful)
Don't worry about that. If Debian dares to mimic Aqua as a theme for their Deskstop, Apple will sue them over look and feel. Jobs is good at that. However, the HIG provided free of charge online will be a good guide for Debian to enhance the usability of their KDE and GNOME desktops for the novice users.
Re:Apple == left handed computers (Score:2)
Start Here: (Score:5, Informative)
http://developer.gnome.org/projects/gup/ut1_repor
I read this about a year ago. It does an *excellent* job of pointing out many of the inconsistencies and gotchas in any given linux desktop situation.
Also (Score:5, Informative)
Re:Start Here: (Score:3, Interesting)
On the first page of the report, I noticed something odd. All of the study's (largely non-hacker) participants rated themselves as "expert" Windows users but utterly helpless at UNIX. Interesting thing is, to a pure end user, CDE/GNOME/KDE aren't too far removed from Windows as far as available tools, etc., go--the main difference is that Windows is flashy and expertly marketed. I think that people, in general, perceive UNIX as "hard" regardless whether it actually is. This psychological barrier is artificial, yet it makes up the biggest obstacle to getting through to most people about UNIX and Linux.
Me thinks Sun and the GNOME foundation need to crank up their respective marketing machines to further dismantle Microsoft's dominance in the global perception about computers and software. Whenever Sun is ready with their GNOME/Linux business PCs, they should get full page ads in the major PHB-oriented and business-oriented periodicals. The word really needs to get out there.
Daunting? (Score:5, Insightful)
Re:Daunting? (Score:3, Insightful)
Sheesh.
Re:Daunting? (Score:2, Interesting)
Just try to replace a videocard - what will Debian with default xdm do?
Re:Daunting? (Score:2, Informative)
Just try to replace a videocard - what will Debian with default xdm do?
Flash x five times, and then start up a text-mode dialog telling you that X seems to be crashing, and politly ask you if you want to reconfigure.
If you select no, it will kindly disable xdm for you, and ask you to enable it, once you've worked out what the problem is.
Maybe redhat copes with this now, but it certainly didn't use to.
Re:Daunting? (Score:3, Interesting)
Shut down, swap network cards, boot up. On init, it says something like "Your network card has changed, do you you want to copy your old settings to the new card?"
Then it does it, and you boot up.
Same thing with video cards/etc.
Re:Daunting? (Score:3, Informative)
Re:Daunting? (Score:3, Insightful)
BINGO!
Why is it that whenever you say you hate the Debian installer, Debheads always assume it's just because you're not accustomed to text interfaces?
I, too, am a Slack user (in the past I've used Mandrake and Caldera for my desktop, but switched once Slack hit 7.1), and I've tried Debian once - with pretty much the same results as you.. I find the installer completely non-intuitive, almost like it was made difficult on purpose.
Why can't the Debian people make an installer that's easy?
Re:Daunting? (Score:3, Informative)
In debian stable (woody), when you are done setting up the base system, tasksel has a few broad catagories.
Afterwords, you can run dselect to individually select or deselect packages, but you aren't required to.
I think debian stable{-1} (potato) was the same way. Never had to install anything older then that.
Re:Daunting? (Score:3, Insightful)
Re:Daunting? (Score:4, Informative)
It took me a while to figure out the exact driver for my sparcstation, and in the end, i had to open the box and do a search on google to know.
This new incentive to push debian into the desktop is "a good thing". Even if it doesnt turn out perfect, it's still a step in the right direction.
Re:Daunting? (Score:2, Informative)
Re:Daunting? (Score:4, Insightful)
Why do people freeze up when they are given a telegraph key to communicate?
Brand-X's difficult to use installer is no more difficult that any other distribution's difficult to use installer.
This is the whole crux of the problem in Human interface. Programmers are NOT good designers of human interfaces. Apple's Human Interface Guidelines said it best. (Back in 1984!) The general sentiment that Apple expressed was this: We, slashdotters, are willing to put up with the most abysmal interfaces, and have done so for so long that we no longer are even capable recognizing a bad interface.
In the 1800's there were widespread predictions that eventually there would be a telegraph on every desk!
But what happened was that a more natural way of communicating (better human interface) came along and appeared on every desk instead.
Re:Daunting? (Score:3, Insightful)
If you have to ask this, you must be a *nix geek.
What's easier to peruse
Which one do you have to sit down and focus on to get an idea of what is going on, and which one can you just skim over and pick up the gist of things?
Now apply this to an installer.
Re:Daunting? (Score:2)
Re:Daunting? (Score:2, Interesting)
The Yellow Dog Linux installation has a graphical installer (using the Blackbox WM) and it requires you to click through about 10 screens. If you live in the USA, most of the default settings are accaptable.
After installing both Mac OS X and Windows 2000 many times, I can honestly say that the YDL installer is the most directly functional and the least obtrusive.
http://www.yellowdoglinux.com/ydl_home.shtml [yellowdoglinux.com]
System Management for User (Score:4, Insightful)
We will try to ensure that software is configured for the most common desktop use. For instance, the regular user account added by default during installation should have permission to play audio and video, print, and manage the system through sudo.
I think giving the root privileges to the user using sudo is a security risk. It will be very easy to wreak a havoc on the system, once you break into the user account.
Re:System Management for User (Score:3, Interesting)
If you're going to deploy this in any kind of serious setting, you'll have admins to set up scripts to remove the glaring security holes in any case.
Re:System Management for User (Score:3, Insightful)
Co-ordination please (Score:5, Insightful)
The OSS community can make a desktop that is better than XP. In fact, all of the bits of the puzzle are already there, it's just that they are in different distributions! (Xandros, SUSE, RedHat, Lindows, Debian...)
When Linux has a reasonable foothold in the desktop market, then go ahead, fight away. But until that day, please share and co-operate. For motivation, imagine Bill Gates giggling to himself and muttering "what a bunch of losers". Works for me.
Re:Co-ordination please (Score:3, Insightful)
Yes, commericial distros have appeared, and their competition is Microsoft (and Sun, and the BSDs, and Apple), but that doesn't mean that the non-comercial distros (a la Debian) need to join in the fight also. They don't have a competition per se because they have no real financial stake in the product. Debian is developed by volunteers and hobbyist who enjoy the work they do. If they didn't get satisfaction from it, they wouldn't be doing it. Do most Debian developers think they are competing with Microsoft? Probably not...
Plus, cooperation between Linux distros is somewhat implicit. Since they are required to provide the source under the GPL, they are already sharing with the community.
I think that the number of distributions is a good thing for the non-comercial distros. As for the commericial distros, well they are already decided to join forces. [unitedlinux.com]
Re:Co-ordination please (Score:5, Insightful)
Well, it depends what you want. If you are happy with Linux having less than 1% of the desktop market, then fine. But there are many people like myself who believe that the world would be a better place if OSS software were much more widely used. And that means the desktop. So, yes, at least from my perspective, Microsoft is the enemy, or if you'd prefer, the competition.
And no, Sun, the BSDs and Apple are not really the competition because they don't own 95% of the marketplace.
Plus, cooperation between Linux distros is somewhat implicit.
Except it doesn't seem to be happening that way. If you look at the kernel, there is a great deal of uniformity between distributions basically because there is a single, widely accepted head-penguin who is doing an excellent job (Linus). I'm sure that there are very few people who would argue that it would be great if we had dozens of incompatible kernals. The desktop space however doesn't have a head-penguin and it really shows.
It's possible to have a high level of co-ordination and co-operation as well as freedom and choice.
Re:Co-ordination please (Score:2)
That statement really made me sit up!
What an interesting idea. The first thing I thought was "Hey, vote for me! I'll be head penguin and really sort all this stuff out." It was just a silly thought of course. But then it got me thinking:
How could we have a head penguin?
What if we had an election, via the net, and voted for a Head Penguin?! I'm serious. Give the Head Penguin a term of say, two years... or maybe just one year to start out to see how it all works. And then continue to have new elections ever one or two years. Who knows what the details of all of this would be like, but maybe a little debate about the possibilities would be useful?
Re:Co-ordination please (Score:2)
I don't think that would be the best way. For it to work the Head Desktop Penguin would need a) to be known to most people in the desktop field b) to be impartial and fair, and c) to be respected. I think the best way would be for the desktop gurus from all the major distributions to discuss it together, and then choose someone from amongst them who would hopefully agree to being Head Desktop Penguin.
Give the Head Penguin a term of say, two years...
I think that due to the nature of the position, the term would be until the Penguin got bored/over-stressed, or given a vote of no confidence by the community.
We already have a good model of how this might function with the Linux kernal. Unfortunately I don't think it is going to happen for the desktop space, which is a shame.
Re:Co-ordination please (Score:3, Interesting)
It's called the Bazzar. Not the Cathederal. It seems to have worked very well so far to create some really good software. Sorry it might not work as quickly as you or I might like.
Remember, your competitioin is Microsoft, they have 95% of the desktop market (or there abouts).
The Bazzar model of development is what will hurt Microsoft. They cannot compete with that.
Choice is what will undermine Microsoft. They do not offer that.
Example: hardware vendors should all work together and cooperate. There should be only one single kind of CD drive, one kind of each size of monitor, one kind of CPU, etc.
The Bazzar seems to work there quite well. Look at hardware prices.
I agree that it would be bad if the open source world spread its resources too thin. But I don't think it is. Even if it were, there is little you or I can do about it. That is just another condition of the Bazzar.
You should be teaming together to fight them, not amongst each other.
I don't think there is any "fighting" going on. I suppose we have seen some bickering amongst KDE and GNOME advocates. (But amongst the developers?)
Re:Co-ordination please (Score:2)
God let's hope not. I run Mandrake, but I hate MCC. It's unwieldy, unintuitive, unstable, and ugly.
The Debian menu system sucks ass too. Great idea in theory, terrible in practice. The upshot of it is you can't use the desktop's own menu editor, or your changes get wiped out next time you install a package. Also, any software that isn't aware of the menu system can't put entries in your menu, or again they will be gone when you install a menu-aware package.
Gnome, KDE, and the other window managers/desktops need to get together and decide on a common location for menu files, desktop icons, and such, then there will be no need for Debian's menu package. Actually, the LSB would have been a great place to specify those things, but it fell way short on standardizing GUI's and such.
Maybe it's too much to ask, but . . . (Score:5, Insightful)
At the least, I'd like to see the Debian compiles updated to i586. (That's the equivalent of a Pentium 1, in non-geek speak.) There are still quite a few of those in use.
Updating the targetted processor architecture would give a significant performance boost to Debian. I mean seriously, nobody is going to run KDE or Gnome on a 386 -- it'd take DAYS just to start a program.
It might also be possible to support multiple processor architectures; eg during installation you get a list of i386, i486, i586, and i686 (386, 486, Pentium 1, More Recent Stuff). Then apt would fetch the appropriate package flavor. Of course, this would require non-trivial amounts of storage space, not to mention all the time needed to re-compile everything.
Comment removed (Score:4, Interesting)
Re:Maybe it's too much to ask, but . . . (Score:2)
Surprisingly quite a bit. I moved my linux workstation over to Gentoo and it runs a lot snappier than my old debian setup. (no hard numbers, simply the feel from day to day use) I have an Athlon 750 and an ancient ATI PCI Rage Pro card so every drop counts.
Re:Maybe it's too much to ask, but . . . (Score:3, Informative)
I suspect that you're probably right about the costs outweighing the benefits. (Sigh.) It's just that when I tried out Gentoo, the difference in execution time was noticeable, and not just in big applications like KDE. I had used custom compiles of KDE and XFree86 under Debian Woody for some time, but the underlying stuff must have slowed it down. Under Gentoo, it takes my machine about 22 seconds to start KDE, whereas under Debian Woody it took about 45. In my book, a 50% decrease in startup time is significant.
Re:Maybe it's too much to ask, but . . . (Score:2, Informative)
glibc, X, and crypto libs would get a benefit from CPU optimization. If nothing else, the order of instructions might be changed to better support a superscalar architecture.
Mandrake accomplishes the glibc by having a /lib/i686 directory with i686 builds of the most intense glibc components. At runtime, the kernel is queried to determine the CPU and based on that, either the i586 builds in /lib or the i686 builds in /lib/i686 are used.
Of course, this doesn't work on VIA processors, as they are mistakenly id'd by the kernel as i686-compatible, when they really aren't.
Re:Maybe it's too much to ask, but . . . (Score:2)
Re:Maybe it's too much to ask, but . . . (Score:3, Informative)
B) compiling for 586 is retarded. The only sytems that benefit from 586 optimizations are 586 systems - 686 systems are architectured differently so that good 586 optimizations don't do much for 686's. Optimizing for 686 would actually give a performance benefit. If you want that, go use Arch Linux or Gentoo.
C) Recompiling all those packages, or keeping both i386 and i686 archives, would be a tremendous amount of work. And, to be honest, 99% of apps don't benefit that much from the optimizations anyways. Recompile your multi-media apps (or use ones that detect the corrent modules at runtime) and install an optimized kernel package, and you should be good.
Re:Maybe it's too much to ask, but . . . (Score:3, Informative)
Not really, do you have any idea of how many platforms Debian is currently autocompiled for? (I've lost count)
Some of these platforms takes days or weeks to compile some packages so there should be pleanty of time to compile the i386 package twice.
No the real issue is that dispite how cool processor specific optimizations sounds, the gains are very limited. I think it is supposed to improve when we switch to gcc-3.2, but it has to be ready for all the Debian platforms before that is attempted.
Re:Maybe it's too much to ask, but . . . (Score:3, Informative)
Re:Maybe it's too much to ask, but . . . (Score:5, Insightful)
> ls -1
libcrypto.so@
libcrypto.so.0.9.6
libssl.so@
libssl.so.0.9.6
There were experiments doing custom glibcs a while back, but there were bugs and it was backed out. I'd like to see that working though.
Great! (Score:4, Interesting)
As soon as Linux is recognized as a userfriendly, easy-to-use desktop with lots of free (as in beer) software by the average user Windows will get into real problems. Such an opening would generate a *huge* increase in the number of users and thus in the interest in supporting Linux from different companies.
Two points to avoid flaming: 1) I know that Linux is only the kernel, simply sed 's/Linux/Gnu\\Linux/' and be happy, 2) "as in beer" is how the average user will see it, my mother don't care for open source, she wants to use it as a tool!
Wrong, wrong, wrong. (Score:2)
As soon as Linux is recognized as a userfriendly, easy-to-use desktop with lots of free (as in beer) software by the average user Windows will get into real problems. Such an opening would generate a *huge* increase in the number of users and thus in the interest in supporting Linux from different companies.
Making Linux easier to install, use, and maintain would be a huge leap forward. However, the VAST majority of end-users do NOT install operating systems. Realize: they can't even install Windows XP, which continues to make installation on x86 easier than previous versions of Windows. How can you expect them to install Linux?
End-users buy computers with OSes pre-installed. That's the key.
Re:Great! (Score:2)
I think it goes a bit deeper though
Wrong focus? (Score:2, Insightful)
-XFree, 4.2 just appeared in unstable
-KDE 3
-Mozilla 1.1
And it's even worse for people using woody without 'proposed-updates' package repository!
The 'testing' distribution is a step in the right direction, but there's a lot more to do that just to focus on Desktop, IMHO.
Re:Wrong focus? (Score:5, Informative)
You also forget that Debian is not a company, but a community. In other words, you cannot dictate what will be done; people will do whatever interests them. It works, it's just that at this point with so many transitions and changes going on, the process has slowed down. Want to sped it up? Fork over some $$$ to a developer. Simple as that.
Re:Wrong focus? (Score:3, Interesting)
I suspect that the reason it's not in testing is because someone found a bug in it that's considered release-critical.
Oh, and GNOME2 (the bits that aren't already in unstable, cf gnome-terminal) is apparently going into unstable this weekend (according to the gtk-gnome list archives). The holdup was transition scripts so that it wouldn't completely throw away all your existing configuration settings from GNOME1 (remember that the two can't coexist cleanly!). They're going to hold it out of testing artificially until these transition scripts have been tested a little more.
Personally I'm extremely glad of that, because I use GNOME to get real work done and I went so far as to artificially downgrade my sid machine to sarge specifically to avoid gnome-terminal and other GNOME2 packages. If GNOME2 had gone into Debian any sooner, I'd have been terribly unhappy.
I'd like to see xft/fontconfig make it into Debian, but the X maintainer has made a good case for holding off on that until X 4.3ish. When you're Red Hat and do one release a year, ongoing changes aren't a problem because you can ignore them until your next release. When you're Debian and do one release a day, ongoing changes hurt!
Stuart.
Re:Wrong focus? (Score:2, Informative)
Re:Wrong focus? (Score:5, Informative)
Debian focusses on whatever the Debian developers care about. One thing Debian developers tend not to care about at all is armchair experts. If you happen to disagree with what we care about, feel free to learn How you can help [debian.org], or to pay for a developer to scratch your particular itch.
We all know that some critical packages are way out of date:
-XFree, 4.2 just appeared in unstable
And excellent prerelease packages have been available from the X Strike Force [debian.org] for months. Not to mention that Debian supports X on 11 architectures [debian.org] rather than just i386.
-KDE 3 Unofficial packages are available; official packages will follow after the gcc transition; see the FAQ [davidpashley.com].
-Mozilla 1.1
Available [debian.org] in unstable and testing, as are recent CVS snapshots.
And it's even worse for people using woody without 'proposed-updates' package repository!
woody is the stable release. Debian takes stability very seriously and the stable release is only updated to fix serious issues (in particular security issues), not to put in new releases of packages. If you want a more up to date system, use testing.
I Have No Problems. (Score:2, Interesting)
Personally I have no problem with Debian's installer. In fact, it is the best I have seen yet. In a desktop situation you can (with the exception of partitioning a disk and one or two other places) just continually hit [ENTER] and come out with a base Debian install.
Which installer is better?
Ximian Setup Tools (Score:4, Interesting)
Anyone...?
Re:Ximian Setup Tools (Score:5, Informative)
Menus (Score:4, Interesting)
Re:Menus (Score:2, Informative)
Mandrake has made the Debian menu system do just that. Perhaps the Debian developers may want to take a look?
there are only two classes of users (Score:2)
The article says "there are only two classes of users: the novice, and the expert"
How about all those novices who think they're experts?
Re:there are only two classes of users (Score:3, Funny)
I believe they are called MCSEs.
Re:there are only two classes of users (Score:2)
Redundancy (Score:2)
WHY the Graphical Debian installer focus? (Score:4, Interesting)
SO WHAT!?! It installed fine.
Some of my minor complaints include:
Ease of install of Xwindows
X installation has always been a bitch for me irregardless of the distribution Linux or BSD. It seems that it's something that always needs tinkering. I did get this going fairly quick after some help from my BSD admin guru--thank the Gods for buddies!
Ease of install of sound
I still don't have sound working, but I haven't given it the one two punch!
Ease of install of APM support
I probably haven't looked in the right place or good documentation doesn't exist. And I'm lazy?
Up to date install documentation
Let's face it, I think that once most people get their configuration working they don't think about giving back to the community. Something that should definately be reconsidered.
Package Manager selection at end of install
Aggravating. I don't want to sit and select then download and install 200M of software after I get it installed, but I DO want an easy way to get back to package management once I hastily exit out if it. I want my cake and I want to eat it too.
Overall though, my Debian install was a pleasant experience.
If I didn't have ~10 years of XP working with *NIX as a user and ~20 years of XP working with computers I certainly wouldn't know where to begin. That's why I think better documentation is certainly in order.
OVERALL, Debian is everything that I would expect it to be for what I consider to be a non-commercial distribution of Linux.
Unnecessary debconf prompts (Score:2, Insightful)
Bravo! Aside from wading through 5 million packages to decide which to install, this has been the worst part of installing Debian for me (which I've done on a number of computers because I LOVE how easy it is to keep my system up to date using dselect). In fact, all the prompts may be even worse.
Here are a few ideas for reducing prompts without causing problems:
1) Make a log of all the prompts that WOULD have been shown so that those who want to can go back and see what else they might have customized.
2) Another reason to make a log of the prompts is in case you accidentally okay one of them and then realize you wish you'd read it more carefully. And it would save you the trouble of writing down anything that it suggests you might want to do later.
3) Give people the option of seeing more or less prompts. Some people may want to see them all. Others may want to only see prompts for things that could make their computer stop working if configured wrong. Others may want more than that, but not every grizzly little detail about configuration files they've never looked at and never will look at.
4) If you really want to get zealous, you could add the ability to make a list of packages that you want to see all prompts for (you'd build it over time) so that you can run on minimal prompts for most things, but for packages that seem to get messed up every time you upgrade them because the default isn't right for you, you get all the prompts.
debian can really do this right (Score:3, Insightful)
well, I'd rather have a solid package management (PM) tool than a simple gui for everything. but now that I've got the solid PM sure go for the solid gui or solid config tools.
I think debian has it right in hitting the important bits first and getting it right before moving on to somthing else. they don't do it half assed.
Re:How funny because (Score:2)
Re:How funny because (Score:3, Informative)
Re:How funny because (Score:2)
Since it's likely that no one will mod up the AC stating the near-obvious, I'll second his comment. You should have a regular cron job backing up your
Re:How funny because (Score:2, Informative)
Check out any recent post from Branden Robinson on debian-devel, most of them cover this, this one [debian.org] for example.
Michael
Re:Isnt Linux Customizable? (Score:5, Informative)
The truth is however that each distro exist to offer you even more customizibilty. You have distros like Slack/Gentoo that many like because they don't include many unnecessary packages and the distro offers you much configurability. Many don't like these distros however because they don't have the time to compile(Gentoo)/configure/install everything the good-old-way or that they just want a distro that is a tad more user-friendly. For those you RedHat/Suse and Mandrake that are distros that are based on a binary package system (Gentoo has ports which downloads the source and compiles it). Each of these have their own "touch" as well.
Mandrake offers many patches/programs to make life easier, so Mandrake is a very popular choice for people that are new to Linux.
RedHat doesn't offer as much as Mandrake in the newbie area, and are a bit more strict on what goes into their kernel and distro. So imo RedHat isn't quite as user-friendly as Mandrake.
SuSe I don't know much about, I know that tthey have a configuration utility that has gotten a lot of positive feedback (YaST isn't it?).
So the choice of distro is just a part of the customization. Part of running Linux is choosing the distro that is right for you.
Re:Isnt Linux Customizable? (Score:5, Insightful)
Let's look at the following targets: server, corporate desktop, home desktop and embedded.
Server: customizability and security is a Good Thing(tm). You don't need GUI tools, they only make the server less stable. You need to have server software installed.
Corporate desktop: you need GUIs. You don't want server software, but office software. Security is important.
This target conflicts with the server target.
Home desktop: *customizability and security* are not important. The home user simply doesn't care and is lazy. He doesn't want to customize everything, and doesn't want to enter a password to install software.
This conflicts with both server and corporate desktop.
Embedded: you want to have as less stuff as possible because disk space and memory are limited. This conflicts with all of the above.
You see? Totally different targets with totally different needs. And this is just an example, the real world is much, MUCH more complicated. Add a graphical installer and you'll piss off the old school power users or users with old hardware. Don't add a graphical installer and you'll piss of the home user. Add an option and you'll confuse the hell out of newbies. Etc. etc.
You can't have one single distro that fits them all, the situations are too different.
Re:Isnt Linux Customizable? (Score:3, Insightful)
Linux "server" distro: Win 2K Enterprise Server
Linux "corporate" distro: Win 2K Workstation or XP Professional
Linux "home" distro: Win 98/ME or XP Home
Linux "embedded": Win CE
Re:Isnt Linux Customizable? (Score:3, Insightful)
Re:Isnt Linux Customizable? (Score:2)
I assume you would want things like "ls" to be part of the core. The "ls" on my Linux box is (c) FSF. Are you saying someone should re-write "ls" just so we don't have FSF stuff in this "core" Linux? There's a saying about cutting off your nose to spite your face that seems appropriate here.
Besides, who is going to develop this "core"? It can't be Linus, he's busy enough with the kernel. I suppose we could use the BSD versions of all the basic commands, but if you wanted to run BSD, why didn't you just run BSD?
I agree that the basis of the system should be standardized (that's what LSB is for), but saying it shouldn't be GNU just so you don't have to hear Stallman ranting about GNU/Linux is a bit silly.
Re:Isnt Linux Customizable? (Score:2)
No, he's just one of a plethora of ungrateful wretches that wishes to obfuscate and denigrate Richard Stallman's rather impressive contribution to the core Linux system (something like 95% of the code that comprises what is traditionally considered the core of a UNIX-like operating system is FSF code), probably because he has a personal dislike for the man, a dislike likely inherited not from personal contact, but from word of mouth and reading the tirades of RMS's detractors.
In other words, he wants to be able to say "see, I shouldn't call it GNU/Linux, since I've stripped out the 95% of the system that was FSF software and cobbled in less-feature-rich BSD versions instead, or did without!" Ignoring, of course, that doing so is, aside from being completely asinine for no real good reason, doesn't reflect the reality for the other 99.9% of the Linux, or if you prefer, GNU/Linux systems that have been deployed.
You are absolutely correct, "silly" is about the kindest way to characterize such nonsense.
Re:Isnt Linux Customizable? (Score:2)
Linux isn't actually Linux at all, it's the Linux kernel AND a whole host of tool, utilities, applications. A distribution is a blend of these OS "parts". There isn't actually any real "magic" about a distribution either - if you want you can gather up your own set of these "parts" even add your own custom built ones and create your own unique "Linux". If it turns out well you might even start distributing it (that's why Linux distributions are call that and not versions).
The configuration you talk about is the Kernel itself, it's what's called a Monolithic Kernel (almost all of the Kernels in general use are Monolithic) but in the case of Linux you can build a Kernel that supports only the features you want or need (before you even get into reworking the Kernel - this is just changing the "make"). Linux's Kernel makes some good pragmatic choices in it's design - it doesn't try and be leading edge, it was designed to be a good conventional Kernel, and later effort was put in to make it portable (it owes much to the wide availability of gcc for it's portability).
Of course people use Linux for a wide range of reasons, and different aspects are important to these groups, this means that they blend their distributions together differently. Mostly this works well, and software is highly portable between different distributions. There are areas that are quite different (package handlng is one example) but mostly there are lots of design choices that are pretty standard. Again, just because you've got a Linux distribution from one group or vender doesn't mean you are forced to accept it as is, you can alter it. The is the very good thing about the GPL - no one can deny you the right to change what you have. You can even rebuild the kernel if you want (in the early days you needed to do this, but not the NEED isn't so common - but it can still improve the system by only building in support for what you have and intend to use)
Of course much of these "rights" might be outside the skill set of those new to Linux, and might not seem important - but it does mean that people will be working away with Linux trying to better solve any problem you can think of. It is that explains how Linux has come so far without huge funding. Lots of poeple see the number of different distributions as a weakness, where as it is actually a demonstration that Linux can be what ever you want it to be: the OS of some monster server or the OS that powers a little PDA or set-top box or anything else.
That's also why Linux users are SO passionate about it - Linux can be whatever you make it. Especially if there's enough of you!
Re:Isnt Linux Customizable? (Score:5, Insightful)
It's actually a case of "too many chefs". No one can agree on what a "great" desktop actually is, since a desktop is so many different things to so many people. The amount of flames I usually receive from people on Slashdot about how feature "X" in environment "X" (or even Windows XP) is stupid and pointless illustrates that well. To one person, it's performance that makes for a great desktop. To another, it's applications. Yet another, eye candy. So far, here's a breakdown of what I've seen:
There are the guys who like to run light with TWM and just use X as a way of having multiple terminals visible at once. I'll bet if there was an option to do this without the overhead of X, they would. As I ocassionally would, myself.
Then there are the folks (like me) who love the eye candy and use X with environments like Gnome or KDE, or just go with a window manager like Enlightenment. These people are the obsessive tweakers who constantly poke and prod and try to get their GUI to be as efficient, customized and sweet looking as possible. To them, the eye-candy/themes, easy customizability and responsiveness of the GUI are what make for a great desktop.
Then there are the utilitarians. They believe that a computer is just a tool. There are also subgroups within this group because some of them want to run light (TWM, FVWM, etc...) and others want an easy to use system that resembles Windows or Macintosh depending on their previous platform (KDE).
Those who use KDE tend to just do the trivial tasks like, browsing the web, working with e-mail, writing docs and balancing their checkbook. They are less concerned about looking cool and more interested in "just working". To them, KDE is the ideal desktop.
The FVWM and TWM crowd that go beyond using X for terms, usually have more specialized needs. They don't need a file manager since they tend to do that from the terminal. They don't need a launcher, since they tend to do that from the terminal too. But they want access to a basic GUI for the apps they run that need it: (GIMP, Mozilla, Netscape, etc...).
There are more groups, but I won't continue. The basic problem is that the desktop is something different for every person. I think the ideal would be an environment that can be as spare as TWM with only xterms/gnome-terms/konsoles, as feature filled as a normal GNOME or KDE environment, and as beautiful as Enlightnment. Upon installation of the environment, the user picks "Thin, Moderate, Full, Custom" and gets what they want. If KDE or GNOME could be made to do this easily and on-the-fly, that would be wonderful.
For instance, if a window manager like Sawfish or KWM could tell it was running with the gnome-panel or kicker, it would revert to a TWM mode. In this mode it would minimize apps to icons on the root window and place a simple right click menu (like TWM's) on the desktop to run a basic set of apps (terms specifically). I think the people who want the lightness of TWM, but maybe like more modern looking widgets and their accompanying functions would really appreciate this.
The ability to switch between Thin, Moderate, Full and Custom, on the fly would be great too. A laptop user could just toggle to Full mode when plugged in and toggle to Light (TWM-like) mode when on battery.
Suspendable X session would be nice too... You suspend the X session and go to runlevel 3 to do deep work that needs all the horsepower it can get. Or you suspend and get back to XDM/GDM/KDM and log in as a different user to do other tasks. (Great for a multiuser machine at home or work) Then you resume your X session with all apps running. This would be kind of interdependent on both the environment and X.
Whatever the case, I think the "Desktop Distro" that will win dominance would be the distro that addresses these issues and does it in a simple way for "Joe User" as well as more complex ways for "Joe Power User" and "J03 G33k G0d".
Re:Isnt Linux Customizable? (Score:3, Insightful)
Re:Once again, the point is missed (Score:2)
However, we dont need all distros to reach there. Thats the beauty of OSS, chose what you like. So currently we have mandrake, redhat approaching the novice user, we have debian for the masters. But requiring some aptitude is actually a good thing. It gives a real novice a great kick to graduate to slightly expirience. The proof is in haloween docs, where the M$ guy says "It was fun, even addictive" . Making ppp work by giving a noauth option may seem rudimentry to a slashdot crowd, but to someone totally unaware of unix world fixing such a problem will provide a great ego kick. That I did it feeling.
So IMHO, the desktops should require some expertise, albeit little, but lots of documentation. Think this way, what would a novice choose, some illegal exception at blah blah blah and then crash, or neat messages telling exactly what went wrong.Look in the automobile world, nowdays cars require expertise with so many gadgets, but the current generation goes for it, coz it kind of gives them with geekiness. Dont forget that a person who can program his HI-FI audio has an aptitude for *tinkering*. Its just the presentation. Give the user the power, make his/her learning curve an enlightling experience.
Believe me, it will work, its the right mantra
Re:Once again, the point is missed (Score:2)
Re:How about making the apt sources... (Score:2)
In any case, the "gnome" metapackage only suggests "gnome-office", it does not depend on it.
Debian will not even suggest you install gnome-office if you just install sawfish.
Get your terminology straight.
Re:How about making the apt sources... (Score:3, Interesting)
In Debian, there is kdebase. It only installs the essentials for KDE. I think it's reasonable to assume that if someone asks for kde, they'd want most of the stuff kde.org [kde.org] offers.
Actually, since I started using Debian again, I've been fairly impressed by the work that goes into just pulling in what you actually need.
Re:Debian dosent support usb mice. (Score:3)
(* Ask The Fine Mailing-list)
(** Fill A Bug Report)
Re:Debian (Score:2)
Too bad you'll have to wait two years after mouse gloves are on the market before there's a Linux driver for them stable enough for the Debian developers to deem it worthy of inclusion into their distro.
Re:X-Windows? Really? (Score:3, Interesting)
Yeah sure, you claim that X* will never cut it in the desktop world. Why? OK, network transparency is not imporant to you, but it doesn't hurt you either! What's so wrong with X that it needs to be replaced?
It's got support for lots of video cards, 2D graphics, 3D hardware acceleration (OpenGL/DRI), hardware YUV acceleration (XVideo), video mode switching (XVidMode), translucent rendering (Xrender), antialiasing (Xft), and soon root window resizing (RandR) and perhaps translucent windows.
And most important of all: X is designed to be extensible!
All the other 'replacements' are big jokes. DirectFB/GGI/Berlin don't support nearly as many video cards and aren't nearly as mature.
* It's the X window System, or just X! NOT "X-windows", "X-window", "X windows" or "X window".
Re:X-Windows? Really? (Score:4, Informative)
What is so wrong with X Windows? If anything, it should be refined to smooth out things people complain about. I'd hate to throw out X's abstractions (client-server; layered architecture: server, window manager, applications) in favor of something new and flashy but architecturally neutered.
I think the fundamental concepts behind X Windows are sound. If there are implementation issues, address those before trying to reinvent everything badly.
Re:X-Windows? Really? (Score:3, Interesting)
X-Windows is going to totally kick ass in the desktop world. Network transparency is the BIGGEST deal ever. Why?
You can run thin clients - for corporates this is a good thing, as it means they get control of their own machines again. Thin client setups are easier to administer, and can work out cheaper than having thousands of breakable PCs that have to be upgraded by hand every few years.
People can try out Linux with only an X server. No really, I've done this several times for people on IRC - they install a Windows X server and I launch a few apps to them to play with. It's a first taste, and often gets them interested. It's how I started
People can "swap" apps between machines on home networks. Well, OK, right now xmove is slightly broken, but the XFree developers are definately considering repairing it and then integrating it into X. Let's say you have 3 computers. One of them has a music setup, as a member of your family is a musician. Another is in the lounge, and another is in a bedroom. You're working on a document in the study, when your eldest daughter wants to come in and write some bangin' tuneage. No problem, you just click in the control button in the window and "Move screen" it to the computer in the lounge, where you can talk to your wife at the same time. A few hours later, she wants to watch TV, so you move again to the bedroom. No need to restart the app. This is easy with X, near impossible with anything else.
You can have logins within a window. XNest lets you do this easily. Can Windows or MacOS do this? No, I think not.
Dude, X-Windows is going to cut up the desktop market. Network transparency is useful to everybody, it's just we've never really had it before so nobody can think of why it's useful to them.