Linked by Eugenia Loli on Thu 16th Jan 2003 04:55 UTC, submitted by Timothy R. Butler
Mandriva, Mandrake, Lycoris "Linux's dreams for the desktop died today with Mandrake's bankruptcy filling. Yes, it was a worthy cause, and we fought hard, but now it is time to admit it -- it is over. No more. Finished. Done. Kaput. GNU/Linux's true place is on the server, and its time for everyone to recognize that. Is everyone with me? Yeah, right. Just to be clear, I don't agree with a single word I just said..." Read the full commentary at OfB.
Order by: Score:
Misleading lead
by sculpin on Thu 16th Jan 2003 05:09 UTC

I hope people take the trouble to acutally read the piece, which isn't as foolish as its lead suggests.

The article's close is much more indicative of the overall argument: "The GNU/Linux desktop is far from dead, and so is Mandrake Linux. If only the naysayers could learn that."

The writer should remember that most people never read beyond the first paragraph ....

I expect there'll be more fallout in the Linux distribution business. I've often wondered, given an open hardware platform, if the compatibility and consistency desires of users creates market pressure to accept software monopoly conditions in order for them to be satisfied.

OSS is, in theory, in a position to get around this by provided a open software solution to match the open hardware it runs on, but while there's a working monopoly to satisfy the market it's going to very hard on the companies that pioneer it's entry in the market and it may be, that even there monopoly conditions are still "desirable" in order to meet the demands.

But I'm no economist.

Ironic
by doomicon on Thu 16th Jan 2003 05:33 UTC

I JUST Mandrake 9 on my this system about 15 minutes ago. I was running RedHat 8. So far I like it better as a workstation. Xine and other multimedia installed flawlessly, while I had problems under RedHat 8.0. I have to admit, I still prefer up2date over urpmi.

RE: Ironic
by doomicon on Thu 16th Jan 2003 05:34 UTC

I JUST.. should be I JUST INSTALLED :-)

I think he's wrong
by Kin on Thu 16th Jan 2003 05:39 UTC

Linux _isn't_ for the desktop. As a desktop OS, it seriously stinks unless you're a computer techny type.

RE: I think he's wrong
by Eugenia on Thu 16th Jan 2003 05:42 UTC

I will have to agree with Kin. Linux is not there yet. In ease of use, usability and integration, Windows, Mac/OSX and BeOS, all beat any Unix/Linux which is based on X11 which just sits on top of the OS, without any real integration with the underlying system.

Why desktop when workstation is still tough?
by GetOutofHere on Thu 16th Jan 2003 05:47 UTC

Why are we focusing on desktop when its still not ready as a workstation. I been spending 5 years trying to install GNU/Linux as a workstation.

Besides what do desktop users need anyways?
browser, chat, office suite, movies, music, and games.
except for games, there's no motivation for programmers.

On the other hand, companies can be profitable in the workstation and server area where they always need service and support. Companies like redhat and suse are heading in this direction.

If any distro is foolish to follow in Mandrakes direction, Chapter 11 waits for you. Let distros like Debian and other non-profit organizations deal with desktop users.

RE: They think I'm wrong.
by Timothy R. Butler on Thu 16th Jan 2003 05:51 UTC

IMO, it doesn't matter if the GUI is some how tied to the core of the system. So long as you can adjust the core settings from the GUI, it seems better to have less integration in this respect. If GNU/Linux's GUI gets messed up, you still have a full system to work with trying to fix things. On the other hand, Windows (for example) leaves you with a sizably weaker shell. I think that's the beauty of UNIX -- the interface is seperate from the functionality, thus no matter what interface works best for you, you have the functionality. Imagine in the future, when this perhaps means a computer designed to take purely voice input, for example.

>Let distros like Debian and other non-profit organizations deal with desktop users.

I do not believe that non-profit distros can do any good or bad on the desktop. The problem with these projects is that they are "recruiting" some people who do have a clue about UIs and usability, and when these people are suggesting changes, the geek developers don't give a sh*t about making changes. Non-profit distros are run by geek devs who develop for fun and they don't listen to (almost) anyone. This is why, IF linux might make it to the desktop one day, it will be a corporation who will do most of the work.

Debian for desktop
by Richard Fillion on Thu 16th Jan 2003 05:55 UTC

"Let Debian and other non-profit organizations deal with the desktop"

While i'm all for debian based desktop aimed linux distros, i would _really_ get mad if they went and aimed Debian itself at the Desktop. I like the fact that it doesnt try to do things for me, it lets me tell it to do things, and it just does them (well).

RE: They think I'm wrong.
by Eugenia on Thu 16th Jan 2003 05:56 UTC

>IMO, it doesn't matter if the GUI is some how tied to the core of the system

Integration is just ONE of the many bulletpoints when characterizing an OS as "desktop-ready".

>On the other hand, Windows (for example) leaves you with a sizably weaker shell.

MacOSX and BeOS does not though. They have bash and tsch. And they are both much better than Linux in almost every way regarding the desktop.

Thinking the wrong way
by Bryce Groff on Thu 16th Jan 2003 05:57 UTC

Everyone goes about thinking that "linux" is or isnt made for the desktop "linux" by itself does nothing. X11 is what allows you to make a desktop and the truth is, is that it is deffinatly not ready for the dumb users. Its terribly complex and cryptic and as old as snot. If linux wants to really go into the desktop market it could but someone will have to be smart enough to ditch X and write there own windowing system. Apple had the right idea and the wrong hardware.

Of course, he's right ...
by Darius on Thu 16th Jan 2003 05:59 UTC

Linux on the desktop is like disco music ... there are too many fanatics who won't just let it rest in peace. I'm not saying Linux on the desktop (asin being a real contender) is dead now that Mandrake is in serious trouble - I'm saying it was dead before it ever got started.
Yes, Apple took the Unix operating system and made it accessible to Joe User, but a parallel to Linux it is not. When you install OSX, you don't choose between 11 window managers, 5 office suites, 3,000,000,000 console-based text editors, and 11,000,000,000 command shells. You can't download OSX off the Internet as an ISO for free, and CERTAINLY you cannot download the source along with it. And most important of all, Apple did what it did on hardware in which they hold a monopoly, and they did it WITHOUT X11.

RE:Thinking the wrong way
by Eugenia on Thu 16th Jan 2003 06:04 UTC

Well said Darius.

> If linux wants to really go into the desktop market it could but someone will have to be smart enough to ditch X and write there own windowing system.

Actually, all the system-oriented parts of the GNU/Linux system need to be more friendly, not just X. For example, when you can't easily install new drivers, or when the kernel is breaking compatibility and drivers need to be re-compiled, all that is a pain in the a$$ for the Joe User. It is not just the X/KDE/Gnome that needs refinement, it is all the system-oriented places of the OS.

RE: desktop
by GetOutofHere on Thu 16th Jan 2003 06:09 UTC

>>>someone will have to be smart enough to ditch X and write there own windowing system.

A smart person will not create a new windowing system for free iso downloads. LOL

RE: Thinking the wrong way
by Eugenia on Thu 16th Jan 2003 06:11 UTC

> someone will have to be smart enough to ditch X and write there own windowing system

Read JoelOnSoftware please. You never rewrite such big chunks of code. You refine them. No matter how bad might or might not be. You never rewrite big code. You fix it and patch it. ;)

I think he's wrng too.
by Tristan on Thu 16th Jan 2003 06:20 UTC

Rather, Mr. Torvalds started GNU/Linux for his desktop computer.

Linus started Linux. If we are going to use the term 'GNU/Linux', we must remember that the project was started well before Linux was.

Is the fact that its NT kernel is the "heart" of most Windows servers an indication that XP isn't intended to be a desktop operating system either?

NT has its origins in OS/2. OS/2 was intended to replace DOS to be the operating system of the masses—including desktops.

That aside, his point is a lot better. I'm quite happy using Linux as my desktop. Sure, there are some things that I don't like, but I'm missing more when I run Windows.

RE: I think he's wrng too.
by Eugenia on Thu 16th Jan 2003 06:22 UTC

>NT has its origins in OS/2.

This is false. Win9x has its origins to OS/2.
NT has its origins on VMS. It was in fact designed by the same person who did VMS.

Re: RE: Thinking the wrong way
by Anonymous on Thu 16th Jan 2003 07:23 UTC

"""Read JoelOnSoftware please. You never rewrite such big chunks of code. You refine them. No matter how bad might or might not be. You never rewrite big code. You fix it and patch it. ;) """
That's amazingly not true. To name an example Sun rewrote a huge chunk of the SunOS kernel, and that is probably the best thing they could've done. While Joel has a point, it's not as blanket as some people make it out to be.

Do you think
by Ludo on Thu 16th Jan 2003 08:06 UTC

Think you that Windows would have been popular with the users if it had been in text mode? No, Linux also. Why Linux would be it chained to brand image of his older brother Unix?

We find basic users who are not afraid of using Linux. It is grace in its aspect of Graphical Workstation : The system is easier of access for all.

The more Linux is known users, the more it enters in corporates as station for development or local web server. It is important also.

Sorry for my english ;)

NT/OS2
by ts on Thu 16th Jan 2003 08:52 UTC


>NT has its origins in OS/2.

This is false. Win9x has its origins to OS/2.
NT has its origins on VMS. It was in fact designed by the same person who did VMS


Why do you think older versions of NT can read OS/2 filesystems (HPFS)?

Ah man
by Mike Hearn on Thu 16th Jan 2003 09:43 UTC

Please. Why are people so short sighted sometimes, especially you Eugenia. No, as of 16th January 2003 Linux on the desktop is not as good as Windows or OS X (for most people) but nonetheless it's charging forward at utterly insane speeds.

I think it's not going to be too long now before it has basically caught up with Windows, and shortly after MacOS (though actually I find Windows far easier to use than MacOS).

The X11 thing is entirely bogus. X11 is a graphics API, it's the equivalent of Quartz or the GDI. The Windows GDI is far inferior in terms of functionality and even speed, yet I never here people bitching about how Windows doesn't feel properly "integrated" because it's based on the GDI.

I think people blame X because it makes it very easy to build alternative user interfaces. I mean, so what? People who say nobody needs to choose between lots of window managers, terminals etc clearly haven't actually used Linux within the last 6 months. Go download Redhat 8 and see. You don't get asked any scary questions and you boot into a beautiful desktop with all the hard questions answered for you. Yes, you can change the defaults, but new users would be fine. X doesn't even come into it.

Of course, people are right when they talk about integration. There are still missing bits, bits that are "nder construction". Menus, themes, fonts, software installation ;) all these things are being worked on and are really pretty critical to having a desktop OS, so really desktop Linux is still in beta.

Now on the other hand, the corporate desktop and workstation markets are less demanding in terms of ease of use and slick integration, and are more demanding in terms of power, flexibility and licensing costs, so it makes sense that Linux is targetting these areas first (gnome is working towards the corp desktop first for instance, as is Redhat, as are SuSE). Redhat are aiming for the end of 2003. After the corp desktop, comes the home desktop (by this point hopefully Wine will be easy to use enough for normal people).

People having to recompile drivers is a two-sided problem, it's only really an issue for closed source drivers (esp the nVidia drivers). If they were open source then they'd be a part of the kernel and would get upgraded as part of the distro. On the other hand, there are some (Very few) driver makers who cannot release the code due to 3rd party licenses.

hear ye! hear ye!
by Solar on Thu 16th Jan 2003 09:48 UTC

@ Eugenia:

> The problem with these projects is that they
> are "recruiting" some people who do have a clue
> about UIs and usability, and when these people
> are suggesting changes, the geek developers don't
> give a sh*t about making changes.

Hear ye! Hear ye!

From both an UI and a developer standpoint, the whole Window Manager concept is seriously broken. A developer just cannot create a "standard" UI since there is no such beast, and an user simply cannot expect that other box to function like his own because Linux simply does not define a standard you could fall back to.

Then again:

@ Eugenia

> Read JoelOnSoftware please. You never rewrite such big
> chunks of code. You refine them. No matter how bad might or might not be.

I disagree with the "no matter how bad" part. If things are FUBAR and by their very concept and design the wrong thing to do...

I thought twice and thrice and yet again before I decided to do just the thing Joel so strongly advises against (and I *do* respect Joel a big deal): Creating my own OS *from scratch*, kernel and everything.

I still keep gettin Flak'ed for that, but a minute or two of calm reasoning usually solves the issue. Surprisingly enough, the more people know about Linux internals, the easier they tend to agree with my POV, even if they consider Linux to be quite fine for themselves.

Now, if only that minute or two of calm reasoning could be made into a single generic document... ;-)

RE: NT/OS2
by GetOutofHere on Thu 16th Jan 2003 09:50 UTC

>>>Why do you think older versions of NT can read OS/2 filesystems (HPFS)?

OS/2 subsystem runs on top of win32 subsystem, which basically emulates the OS/2 operating system. That's why you can run OS/2 apps, same goes for POSIX apps.

Read Operating System Concepts.
Pretty good book on OS.


Jenny is right, NT has it's origins in VMS! I am not sure about Win95 and OS/2, though. OS/2 has both win16 and win32 processes running in separate virtual machines and memory regions (has full memory protection), and it has a real filesystem, not the Micros~1 joke from Win95. I really fail to see any common points between Win95 and OS/2.

Not to say that OS/2 Warp was called the Win95 killer (as it was expected to beat Win95 on features and release date). So IBM intended OS/2 -against- Win95, not to help Win95.

Linux on the desktop died years ago
by riggits on Thu 16th Jan 2003 10:09 UTC

In case nobody's noticed, Linux on the desktop has always been pretty much a non-starter thanks to Xfree86. Sluggish by design, and intended for server use alone, Xfree is the single biggest stumbling block for those who would replace Windows with a high-performance free software alternative.
Ever noticed how clunky KDE is, or GNOME, or any other window manager? Do you actually like your entire desktop to refresh instead of just those "dirty" parts of a window, and do you actually enjoy the latency caused by this monstrous single-threaded application layer between the kernel and the window manager?
Windows doesn't suffer from a bloated GDI layer, and neither should Linux. I like the OpenBEOS idea. Linux will never be adopted by those who care more for graphical performance than how "free" their OS is .. and remember that Windows is "free" in any case to most home users.

free windows everyone
by GetOutofHere on Thu 16th Jan 2003 10:30 UTC

you know it wont be free anymore.

>>>application layer between the kernel and the window manager?

X and the kernel are independent. I think this is also true for the windows operating system.

But your right, that X is a bitch.

Linux on the Desktop,
by Travis on Thu 16th Jan 2003 10:31 UTC

Its amazing, the arguments presented in the comments above, is why a lot of BeOS users have been complaining, (and got bashed about) for a very long time now.

Linux developers, If you want an OpenSource Desktop oriented operating system, with clean and clear standards, thats Posix compliant, OpenBeOS is waiting for you.

dead as a desk top???
by The Pickle Man on Thu 16th Jan 2003 10:43 UTC

Yes Mandrake was pretty - but slooow - not much better than windows. In fact in too many ways it is too much like Windows. If you hope cut out a market share you must be different from the competition. Mandrake has been very busy trying to build a "Windows like Linux". While Mandrake and Lindows will get some play at the beginning of the move to Linux, but the power players will be those Distros that build a fast and usable computer OS.

One of the real advantages of Linux may well come to view from Mandrake if the firm does go under. In that Its open code will still be there for new firms to pick-up and advance on. When proprietary software firms close they often take their code into oblivion with them.

To bemoan the end of Desk Top Linux form the failure of one firm is to Miss the real strength of Linux. Many, people working out their own solutions while communicating with all others in the field. Many other firms are developing Desk Top Linux Distros as well, some I feel have a better handle on it or at least part of the problem. Remember that Linux is less than 14 years old and that most GUI desk top efforts
has only been making a real effort for under 2 years now. Remember Bill Gates spent many years trying to get Windows working unless you consider versions 1 & 2 workable.

The fact is that Linux is making very rapid success in this manner, I consider most Distros I have used to be head and shoulders above Windows 3.1 already in usability and are in the act of blowing past Windows 9x and closing fast on XP. True Linux will never be perfect, but then again I know the short commings of Windows quit well in that I spend my most of my working time trying to get some version of Windows to work at least well enough to get the job at hand done. When I show up and some needs help with Linux feel like jumping for joy - you can bet it will fix easier, work better after the fix, and stay fixed.

Give Linux another year and Not only will it will have many Distros running better and faster than XP. In this time it will gain a very good set of applications for it.

OpenBeOS?
by GetOutofHere on Thu 16th Jan 2003 10:50 UTC

what's that?

doesnt sound sexy enough for me.

HAHAHAHAHAHA.

let it die

MAC on x86 sounds more sexy

open arms
by Chris Simmons on Thu 16th Jan 2003 10:52 UTC

I would like to welcome any linux coder into the BeOS Community, with open arms, right here and now.

We've always been at the short end of the stick, when it comes to recognition of having a real desktop OS, but I think it's time to call attention to the things that make BeOS so great.

openBeOS and it's future ilk will pave the way for how things -should- be done on the desktop, bar no other OS except for Apple's MacOS X, but that is a given. The concepts and thinking processes that have evolved into what we know as BeOS R5 Pro/PE will only serve as a foundation for things to come in openBeOS, and related projects. Think projects like Glass Elevator are already working on ideas that will blow you away, in terms of elegance, professionalism, and ease of use issues.

With that in mind, I would like to ask any Linux coder willing to accept a challenge, who has a desire to help create a true desktop OS, to step forward, and let themselves be heard on how they can help.

We welcome you. Really. There is no animosity between the BeOS Community and the linux community. None. We are all fighting the good fight, namely, the fight for the right to more choices in computing. The right to have a real choice other than that which is handed to you by the computer manufacturers and their bundled deals that limit our freedom(s).

I think it's time to put aside petty differences, and work together on the things that each computing area is excellent at. Whether that is in the server space, then fine, continue that work. However, if you wish to help create a true desktop OS, then please, by all means, join with the BeOS Community, and help out.

No one ever said that you can't belong to more than one community at a time. Choice is good.

Thank you for your time.

Sincerely,
Chris Simmons,
Avid BeOS User.
The BeOSJournal.

Linux on the desktop, an opinion. :)
by Devon on Thu 16th Jan 2003 11:38 UTC

There is only one reason I can see anymore why Linux is not a common desktop choice. The reason? Because Linux is not a common desktop choice!!! Its a catch-22.

Linux cannot be a truly common desktop choice until it recives at least a moderate level of support from hardware and software makers. It doesn't need a Windows level of support, just a Mac level of support will do. Unfortunatly, it won't get that level of support easily. It will have to become a common desktop choice in order to get it. Even just 5-10% of the desktop market might be enough, but its still got to get there on its own somehow, without the support that is necessary for it to get there. Sucks to be Linux. Thankfully however, it doesn't suck to USE Linux, far from it! ;)

Any other reasons I have seen why Linux is not ready for the desktop just stem from personal preference, and really have little to do with Linux's actual viability.

Just another opinion
by Tyr on Thu 16th Jan 2003 11:38 UTC

Linux on the desktop could happen at any time, it just takes someone willing to do the work. You don't even have to replace X - hell my Redhat 8 runs just fine even on my old systems.
What linux needs are 3 things:
1) integration : all the software and subsytems need to work together, now it's more a kit-job than a real os.
2) abstraction : confusing things like the filesystem, sytem libraries and configuration files that a user normally doesn't (shouldn't) need to know about should be hidden.
3) easy installation of hard & software : no dependancy hell, no recompiles, definately NO kernel-recompiles. Software-installation should be universal, not distro dependant

Unfortunately these are also things linux-geeks don't really miss, they are hard to code and you will get exactly zero credit from the average enduser. So it will probably never happen, but never say that it can't be done.

eugenia
by Karl Yollner on Thu 16th Jan 2003 11:51 UTC

Eugenia,

Apple starting with the Lisa shifted their emphasis from computers-for-hackers to computers-for-consumers. If it had not been for the hackers, amateur enthusiasts and long-haired computer types(including the founders of Apple) neither you or I would be using PC's and making comments about the plethora of Operating Systems. The first time I saw a Lisa, I was blown away, I saw this monitor in a big case, a keyboard, and some strange thing with one button on it. I grasped the thing and to my surprise I saw that something on the screen moved as I moved this thing. I pressed the button and I was drawing on the screen-wow. Now anyone who used home computers back then (1982) can realize what kind of change this represented.

Lisa, likes its Mac successors, was a graphical operating system. Everything was implemented at the GUI level. This appears to many to be a form of intergration, and in one sense it is. But as cool as this development was, it also heralded a time when the user was confronted by software which made all choices for them, in advance. Lisa, likes the later Macs, did not represent a integrated system/GUI- for their was no system independent of the GUI.

For comuputer consumers, ie. those who are primarily consumers of digitial products, this step, taken with the Lisa, was a massive step forward. For those who had enabled this technology to be- it was a bane, for they found themselves on the outside looking in.

Apple systematically shut people out- first their was their pricing, back then only the rich could afford one- then their was their propoetary hardware-incompatible with the rest of the world, much more expensive- and then of course the fact that computer users where forced into becoming consumers- becuase they had no control over the system, excepting for those controls which the programmers felt inclined to implement for them.

As much as I was impressed by the Lisa, and later by Macs, I also relaized intuitively that these machines were not for me. My ability to control the system, is what enables me to understand the system, and this is what makes computers valuable for me. the newest MAC OS X takes a step in the right direction, but on the wrong foot. Now we cqn talk about system/GUI integration- the "system" and the the GUI are now distinct entities which are tied together. The *nix-base of the new MAC OS X was dead before it got started. It is not compatible enough with the other BSD's to really beneift from their software innovations and it is too inompatible with linux, which is grwoing by leaps and bounds and which is displacing Apples meager market share.

The idea isn't bad, but if they had based it upon linux or one of the BSD's their system might have had a future. As it stands the MAC OS X represents a stunning GUI based upon half-aborted *nix implementation, - Mac users don't work in *nix, and don't write *nix code, they consume and produce digitial products- the mach kernel outside of the context of the MAC is so underwhelming that it will never draw any large number of developers. There is no future here.

BEOS- I have tried to get it up and runningon three different machines- no success worth mentioning. I would love to see the spirit of BEOS succeed, although BEOS itself is long since dead. HAts off to those who develop it- if they manage to ever get a version which works with any real world computers-ie. supporting less than 1% of computer hardware means that it remains a niche OS for those who happen to have computers using exactly that 1% of supported hardware-I will download it and try it out- I would love to know what all the hype is about.

BEOS is 90% vaporware, it is an unfulfilled promise, not unlike unrequited love, romantic and alluring, always teasing and tempting. Such is beautiful, and has a value in its own right, but cannot be compared against things which actually are. Your comparing other Operating Systems to BEOS is silly at best and distortive at worst. BEOS was never realized sufficiently in order to serve as a STANDARD OF MEASURE, and therefore your comparisons remain unfounded.


Linux-and by the way which implementation are you refering to- is already used as a desktop system by more people than have ever even heard of BEOS. Is it perfect, by no means. Does it still need to be improved, of course. But Linux works, and it works on a plethora of hardware and is usefull in untold numbers of contexts. Linux is not primarily aimed at computer consumers-it is aimed primarily at developers, network administators, enthusiast and hackers. The primary thing which limits linux's role as a desktop machine are the propietary fonts, codecs, formats and hardware-which the consumer expects to be supported.

If the fonts were freely available, without patent issues, or sperately purchaseable for reasonable price, if the codecs were open and freely available, if developers of office-software would use open formats, if the hardware developers would reveal their code,--all of these if's, all of which are determinate for the paradigm of computer consumption, all of which exist in their current un-open form in order to maintain existig monopolies and prohibit competition based on merit, -leave linux in a state of suspended implementation- any distro which correctly implements these things (ie. fonts,codecs and propietary (closed-source)drivers) is subject to being sued. This leaves the end-user to iplement these things as seen fit-and this, alone, is what makes linux "not ready" for the desktop. Luckily these are primarily issues for computer consumers. Linux carries on the tradition of computing which ushered forth the whole consumer paradigm- which made this pardigm possible, linux is not subject to this paradigm and cannot be judged by it. If all of the above listed issues were resolved, linux would already be THE desktop system, rapidly displacing Microsoft.

However these issues will probably never be full resolved and when the computer consumer paradigm has been exhausted, linux will still be plugging away.....



Personally
by rajan r on Thu 16th Jan 2003 12:09 UTC

I think Linux for the desktop isn't dead. but Mandrake is good as dead. Until they can find a business model that actually works, they are probably better off as a non-profit organization.

BeOS = Vaporware
by Jack Burton on Thu 16th Jan 2003 12:54 UTC

Oh, Dear... I was running vaporware on my pc and I didn't know...

Keep the packages top a minimum!
by civad on Thu 16th Jan 2003 12:58 UTC

"When you install OSX, you don't choose between 11 window managers, 5 office suites, 3,000,000,000 console-based text editors, and 11,000,000,000 command shells."

Very true. Though I agree that the users are free to install/use whatever application they intend to use, simply bundling a zillion packages together does not make sense.

Maybe someday one could 'customize' their distor by choosing a list of apps they want BEFORE downloading the iso. A customized iso could then be ready for download for them.

Frankly speaking, I wouldnt mind paying for such a system ;)

Destined for desktop
by ryan on Thu 16th Jan 2003 13:22 UTC

with MS screwing up the way they are and linux improving, linux is assured a place on the desktop. Its going to take some time but it will happen. I think it'll grow slowly for the next five years and then really blossom after that.

I think it will become particularly popular in the business space. MS's dominance won't last, nothing ever does.

Re: Keep the packages top a minimum!
by EcHo2K on Thu 16th Jan 2003 13:30 UTC

ok you already have that, it's called Portage (the base for gentoo), i'm thinking of rolling out a "minimal" distribution with a basic system, then compile ALL the portage tree and let portage fetch the binaries instead of the source/compile (that is a very cool thing, but it's a problem to sell a system that has to stay unused for 2 days beacues you are upgrading KDE...)

X is the problem... suuure.
by Jon Doda on Thu 16th Jan 2003 13:30 UTC

I'm sorry, but anyone who thinks the fact that X is "sluggish" is what keeps Linux off the desktop has their head so far up their ass that I have to wonder what they're breathing. X does present usability issues, but now that we have fontconfig and RandR most of them are in the realm of input configuration. Responsivness is nice, but if that were the biggest problem with Linux GUIs then the battle would already be won.

Linux for the Desktop
by Jason on Thu 16th Jan 2003 13:45 UTC

I don't think Linux has far to go to become viable for the desktop but I do believe it will take two things to make it a realistic option for all users. First the word 'compile' will never used in a solution for any average users hardware/software problem. And second a person should just be able to click on whatever package they have downloaded to be able to install an application without ever having to worry about dependencies or what version of Linux they are running. I'm all for Desktop Linux and am ready to ditch Windows but Linux just isn't there yet for me and many other people.

Linux dead on the desktop?
by Richard James on Thu 16th Jan 2003 13:59 UTC

Of course linux is dead on the desktop. There are no companies like lindows, lycoris, xandros putting money into this area. Linux doesn't continue to improve. There will not be new versions of KDE and Gnome out this year and better ones after those.

Anyone would be blind to think Linux has a desktop chance according to what most people in this thread are saying.

Well stop and think

There are companies investing in Linux and specifically UI advancements.

KDE and gnome will continue to get better.

XFree will continue to get better.

Linux is a desktop baby it grows every year. It is not dead or dying.

X was not designed for servers it was designed to run windowing apps on workstations. X is not slow, it does not look ugly.

If you have a true X vendetta then please tell me you actually use it and have used it for over 5 years. If not it sounds more like you are trolling.

on NT VMS and linux desktop
by MP on Thu 16th Jan 2003 14:07 UTC

To be strict, NT is based on OS Mica. When Cutler's group was kicked out from DEC in 1988 they were working on new project called Prism (hardware) and OS Mica (hmm OS). Gates almost instantly hired Cutler's group. The idea was to challenge UNIX. At first system was called OS/2 NT as it was implementing OS/2 APIs. Of course Mica is based on VMS so there are obvious similarities between VMS and NT.

Now for me the fact that Mandrake is dead seems not to be such a bad thing for linux.
Diversity is good, but if linux want to compete with "standard" then unity is really important. Having two "official" distros is enough (there is only one...ehem Windows or Mac for win or Mac users)

FreeBSD is my workstation (blackbox as WM) of choice and I will not switch to anything else but I had to add to Mozilla's prefs.js a line: user_pref(“general.useragent.override”, “Mozilla/5.0 (compatible; MSIE 6.0; Windows NT 5.1)”); to cheat my bank. Otherwise I would be unable to do on line banking.
Good or bad Windows is everywhere and that makes it a standard. One can fight this standard not by imitating but making something new and different.
I belive that it was really bad idea to mimic Windows GUI on linux (KDE but also Gnome), because Win user will ask why the hell should he use linux version of windows if Windows (MS) is already here with familiar annoyances including menus and program names and so on and so forth.
I belive that the only way to beat the beast is to make something different and better not as good (FROM DESKTOP POINT OF VIEW) as Windows. With KDE or Gnome linux lost the speed, icons are flashy but for how long are you going to look at them?
If programs are as good as in Windows there is still not a point to switch.
Second big thing except having better GUI (for Windows users) is to have linux pre-installed on Dells Gateways Compaqs Toshibas desktops and laptops. I never had problems with installation of new OS (having read requirements beforehand if that was something new) but in general this is a disaster. Even here at OSNEWS I remember an idiot reviewing - I belive - SUSE's 8.1 install. It was really bad. Why? Because even without reading hardware requirements printed on the box he was trying to setup SUSE on the system without enough RAM.
So is it linux installer bad? Nope is as good (almost) as Windows, but Windows users have their OS pre-installed so even if on average the percent if installation failures is similar for both OSes nobody (maybe not exactly nobody) is noticing it in case of Windows as the system is pre-installed.
And the last but easy to improve thing for linux desktop is the sheer number of overlapping programs that are installed by default. For what? To please everybody? Impossible. I think that Mac has now 10 different browsers. All of them are preinstalled? I dont thing so. But if Mac user do not like Chimera or Safari she/he can download and install Opera and get rid of the other two.
In general I thing that linux desktop should stand out from the crowd to win desktop battle instead of mimicking what is already available. I dont know how to do that, as I said I am using blackbox and I am really happy.

Well that was long
MP

X and the Desktop
by Joe P on Thu 16th Jan 2003 14:14 UTC

First off X was designed for the server and not the desktop. The basic concept behind X was to have applications talk to the API which would send the data across the network to the dumb terminal where it would be rendered. The xFree86 team is still supporting this model; even on the desktop. This means that all communications between the user and the application goes through the TCP-IP stack which makes it less responsive.

I aggree that you don't rewrite such a large project from scratch. But, it can still be fixed. First, rewrite the rendering part to be accessed directly from local APIs and bypass the network interface. Second to support running server processes, create an utility to connect to the remote program (the same thing the X emulaters do on Windows).

This would split the xFree86 project into 2 branchs. 1) The Application Server (X applications to Network) and 2) the Client Side (local X applications to the Renderer and connection to the Server)

ryan -- destined for dung heap
by rockwell on Thu 16th Jan 2003 14:29 UTC

//with MS screwing up the way they are //

Uh .. how, exactly? You call $40+ billion in cash reserves (even _after_ the CA settlement) screwing up?

Post something worthwhile .. or provide facts.

Fsck-tard.

X11
by eNTiTy on Thu 16th Jan 2003 14:42 UTC

For all you 'X11 is ugly'/ 'X11 is slow'/ 'X11 is the problem anyway' peeps, read this:
---snip---
[7 January 2003]

Apple, a leading innovator in graphics systems, has announced support for X11 under Mac OS X, offering a complete X Window System implementation based on XFree86, the de facto standard for X11. More information is available at Apple's web site.
---snip---
X11 is THE STANDARD graphics server for ALL Unices and Linux, and it sure ain't slow - it may be hard to use, but it's quite powerfull. And if you ever used well coded X11 apps (e.g. NothingReal/ Apple Shake, SiliconGrail Rayz) - this apps are faster and more responsive compared to Windoze. And for window managers beeing sluggish, check enlightenment or waimea...
So, think again...

Re: ryan -- destined for dung heap
by rajan r on Thu 16th Jan 2003 14:56 UTC

Actually, I have to disagree with you. Microsof have let their guard down in some many areas, the sad part is that nobody even noticed it. Microsoft is more vunerable now than a decade ago, and it isn't because of antitrust nor Linux.

For example. Their most vunerable spot, IMHO, is Office, whom recieved very little competition (in comparison with stuff like Windows, MSN, XBOX, PocketPC, etc.) and therefore a easier target. Sun is the only one targeting it but they aren't doing it hard enough.

Re: ryan -- destined for dung heap
by Croaker on Thu 16th Jan 2003 15:05 UTC

For example. Their most vunerable spot, IMHO, is Office, whom recieved very little competition (in comparison with stuff like Windows

So, Corel Office, Star Office, Apple Works (or whatever Apple's office suit has been named in the past), and whatever Lotus's Office suite was aren't/weren't competition?

The fact is that the longer Office dominates, the deeper it becomes entrenched. Anyone coming into the office software arena is fighting in Microsoft's backyard, since they will have to be compatible with office in addition to competing with Office. Then they have to overcome the fact that people usually get Office pre-installed on their desktops, right beside the OS. This is a point the author of the article overlooks, BTW, when he stated that Linux was the only OS where you got an office suite installed with the OS. Technically, it's true, but consider the fact that the *vast* majority of computer users don't install their own OS, and get systems pre-installed with Office.

XFree Speed
by Nick Slaughter on Thu 16th Jan 2003 15:22 UTC

Even some of the head programmers on XFree admitted it lacks the proper speed to compete now when Linux is slowly getting out on the desktop market for real (largly thanks to Redhat).

Feature wise I have no problems with XFree, especially with RandR coming soon and all the other great things we've seen grow and mature (font rendering, gui toolkits, etc).

XFree is an excellent piece of software, many parts of it have been finetuned for years so rewriting them would be pointless, but it's still slower the Windows/BeOS on the same hardware, is it solvable?

GDI IS in kernel space
by gabou on Thu 16th Jan 2003 15:39 UTC

Someone said

X and the kernel are independent. I think this is also true for the windows operating system.

It is false for windows. Forr all version of NT befoir 3.51, it is true, but for performances resasons, gdi is in the kernel space since. That's why these versions are less robust ( but still good enough for desktop users ).

I like Linux for some things ( it is cool to have a NFS/SAMBA/SSH server on a local network, I can access my computer anywhere in my university ), but for the desktop, it stinks. Really.

On X and the Desktop
by Johnathan Bailes on Thu 16th Jan 2003 15:50 UTC

Look the XFree bashing is silly. X needs some serious tweaking and cleanup. This is true. To blame all the troubles of speed in linux on XFree86 is just wrong, bad wrong.

Window Manager, App programmers, widget class programmers and desktop environment groups all share part of the blame. Honestly, the respective companies left in the linux business need to capture all these people in a room and lock it hard until they come together with some real solutions.

Better user-focused system configuration tools are needed and they need to be integrated either in the top-level menu (RH8) style or into the control center for the desktop like SuSE. Redhat makes some of the best user-focused system configuration tools -- Time/Date config tool is nice, user config is nice, blah..blah but there need to be more of them. I want a damn cron editor like gnome-crontab included.

Distros should stop using slow as heck Mozilla as the default browser and use something like Phoenix for speed's sake especially when Kmail or Evolution is used as the default mailer anyway.

OpenOffice has to port at least part of its look and feel if not its entire project to a widget set that can at least mimick the QT or GTK looks with theme hints. The look and feel differences between the Office suite and the rest of the apps regardless of desktop are too harsh.

All distros including RH have to include tools for easy installation of those vendor-specific drivers for Nvidia, ATI and the installation of MS core fonts that are still available on the web just in other places. Some distros do better in this regard than others. On the fonts issue, a deal with Adobe for some of its font suite bundled in would be a nice coup for one of these desktop distros don't you think? I have the Adobe fonts on my SuSE box. (Our company bought the full font suite package thank you).


I know that Windows people have to download the latest drivers and install them but linux has to be even easier because you get no vendor CD for linux with your hardware. BTW, rpm installation of drivers have caused me very little trouble but I did not try nvidia drivers when RH8 first came out. :->

There has to be wine compatibility layer to ease the transition for corporate use. I know how silly it is. I use OpenOffice and it meets my needs even in a corporate environment but there will always be that one vendor-specific spciality app every corporation will have it so CrossOver Office or at least Codeweavers wine should come with every single linux. Make the deals with the CrossOver folks but it has to happen.

In the end the above steps will help linux on the desktop but linux on the desktop will always be a niche geek thing until three events occur.

1) MS would have to fumble and tick of the corporate base to the point that looking into alternatives like linux becomes a serious part of their IT strategy not just something one manager looks into briefly.

2) MS would have to fumble and tick off the hardware makers to the point they start shipping linux on many of their machines in large numbers.

3) The OSS crowd would have to come up with a killer app or OS function that truly left MS in the dust and scrambling and the companies posed to profit would have to run with the ball as opposed to wallowing in the short term profits too long and letting MS catch up.

All of the above conditions depend on Microsoft fumbling. There are only two conditions that bring down a big Monopoly player. The government breaks them up (this ain't gonna happen) or they fumble really hard more than once and the competition has to capitalize quickly.

Until then linux will be a niche player on the desktop.

Same old, same old
by William Clifford on Thu 16th Jan 2003 16:00 UTC

I've read this same tired old desktop fight way too much. Desktops are matters of personal preference and vary quite widely even between individuals who are using the same platform. We all read recently what some guy had to do to make a Linux desktop for an old lady with limited computing needs.

The story of Mandrake's business problems and what it means for Linux on desktops is bigger than this old fight about the who makes the best environment for general users. I'm disappointed.

Tangentally, I wish the Be and Amiga people would port or clone more stuff to GNU/Linux and BSD systems (not to mention Windows and Mac). If they think they're so cool why not let others try them out?

X
by Mike Hearn on Thu 16th Jan 2003 16:27 UTC

Will people please get a clue before talking about X? Some of the posts here clearly have no idea about its internals AT ALL.

JoeP: "First off X was designed for the server and not the desktop. The basic concept behind X was to have applications talk to the API which would send the data across the network to the dumb terminal where it would be rendered. The xFree86 team is still supporting this model; even on the desktop. This means that all communications between the user and the application goes through the TCP-IP stack which makes it less responsive."

Yes, just like Windows apps talk to the API by sending data into the kernel.

X traffic on the desktop goes via UNIX domain sockets and SHM segments. That means practically no communications latency, it's a direct route via the kernel, and for large amounts of data (the only kind that really slows down inside the kernel) the pixmap data is shared with the X server. The speed of X is just fine, any objective tests will tell you this.

"I agree that you don't rewrite such a large project from scratch. But, it can still be fixed. First, rewrite the rendering part to be accessed directly from local APIs and bypass the network interface."

No, that would gain you almost nothing. For starters you'd need to integrate it into the kernel which would a) make graphics non-optional without a recompile and b) drastically reduce the stability of the system.

The only overhead in X is the dual memory copies, but that doesn't really matter for non-pixmap messages, and pixmap messages use shared memory.

There is currently a bug in the XFree 4.2 scheduler which sometimes starves programs of timeslices, in particular when using opaque resizes or moves. This is a known issue and is due to faulty scheduler rules, NOT any inherant problem with the design. I think this is what people usually refer to by "X is slow", in fact it's simply a bug.

Please, learn something about the X architecture before giving "solutions" which will "fix linux on the desktop". X isn't perfect, but really it's getting very close, and I'd challenge anybody to deny that when they had all the facts I do. X is not the problem, nor is waiting for Microsoft to fumble. It simply needs a lot of work, a lot of polish, which is what is happening.

Xdirectfb
by philicorda on Thu 16th Jan 2003 16:49 UTC

ADVERT

If you have strange, imaginary and misinformed problems with X, or even amusing problems like confusing it with a Window manager, system configuration utility or program installer, there is an answer!

You can use xdirectfb! http://www.directfb.org/xdirectfb.xml

(Disclaimer: Evaluating this product requires actually installing Linux rather than just parroting comments about it you have read elsewhere on the net. Not all video cards supported.)

comments
by Roy on Thu 16th Jan 2003 16:59 UTC

I've seen a lot of drivel and a lot of good points here. Here are my thoughts.

Linux isn't ready for the desktop yet, but it definitely isn't dead either. There are still a lot of pieces missing from the Linux desktop, but it is certainly improving. KDE and Gnome are both improving at an amazing rate. More desktop oriented distributions will probably die, but more will eventually pop up. Success of the Linux companies will have as much to do with growth/money management than a breakthrough product. Mandrake, though having a great product, was poorly managed. Linux won't suddenly takeover the desktop, but I believe that it can slowly gain share.

X11 is not the reason that Linux is having trouble on the desktop. Lack of good driver support and application support (MS Office) is a MUCH bigger problem.

X11 should eventually be replaced (it is definitely showing its age). Here is what I consider the "proper" way. Design a new, better graphics interface. This will only way this will work is with buyin from the major Linux and Unix companies. An X11 compatibility layer should exist. Stop writing applications directly to X11 (this is rarely done now anyway). Port QT, GTK, and other high-level graphics libraries to this new interface. Allow X11 to slowly become irrelevant as fewer apps need it.

Rewriting large pieces of code rarely makes sense from a "business" perspective. In the LONG term is can make sense to rewrite large pieces of code from a development perspective (the "business" problem is that companies can't survive the redevelopment period).

WinNT/2000/XP has roots in VMS, though it had different design goals than VMS and has been modified significantly since its original design (moving graphics stuff into kernel with NT 4.0, I think). Microsoft obviously intended the NT kernel for both server and desktop use. However, there is nothing about the Linux kernel that makes it "bad" for the desktop. Current desktop problems (latency for instance) are being addressed with the 2.6 kernel. Just like X11 is THE problem for the Linux desktop, neither is the kernel.

RE: I think he's wrng too.
by JC on Thu 16th Jan 2003 17:45 UTC

>NT has its origins in OS/2.

>This is false. Win9x has its origins to OS/2.
>NT has its origins on VMS. It was in fact designed by the >same person who did VMS.

This is false. Win9x was still DOS based. OS/2 was not. And if you care to do further research, you can find any number of "History of OS/2" resources which point out that NT was OS/2 based.

RE: I think he's wrng too.
by Anonymous on Thu 16th Jan 2003 17:57 UTC

>>
This is false. Win9x was still DOS based. OS/2 was not. And if you care to do further research, you can find any number of "History of OS/2" resources which point out that NT was OS/2 based.
<<

Incorrect. NT did initially start as a joint project between MS and IBM, yes, and in fact carried the moniker "OS/2 NT" for a while, but any and all form of OS/2 relationship is, and always was, in the form of the OS/2 subsystem present in NT (os2ss.exe). Architecturally NT bears no relationship whatsoever to OS/2. The fact that OS/2 at the time was written almost entirely in assembler was in fact one reason for not basing off it, since one of NT's design goals was portability.

come on people
by Micky Wicky on Thu 16th Jan 2003 18:34 UTC

Stop b*tching about the flaws of Linux and get coding, for G*d's sake!!
I use Linux 90% of the time (the rest is Solaris and a bit of Win2K), and as far as Joe Corporate User is concerned, Win2k is still the way to go, so what? I like Linux as it is and most importantly, how it evolves. I think we should all stop getting other people's code for granted, just because we don't pay for Linux and pay for Windoze when we purchase our PC. To anyone criticising Xfree, I say, "give me another interface that runs on so many OSes, with so many apps, that can be überoptimised by source compilation, and that fulfills just about everyone's needs." Right, now who wants to compete?
The Windows and OS-X interfaces may be slicker, but they're somewhat less flexible. And all three provide almost everything one needs.
Yes Linux is for techies, but think about it. One Techy named SysAdmin can manage dozens of Linux boxen remotely, and the users just get on with what they have to do. Only home users have to worry about new drivers, etc. The desktop is, and always was, a corporate bastion. Home users just adapt to it and think how cool it is to have the same thing at home.

And Another One Speaks
by Ash on Thu 16th Jan 2003 19:18 UTC

I'm am getting pretty tired of people/articles saying 'Linux has made it to the Desktop!' and 'Lets face it, we fell short.' I've been hearing these same statements since 1998 and it's more than old.

Who cares if Linux makes it to the Desktop for the general masses? Look at BSD, it's 'made it' in the form of Apple but I still hear people complaining about BSD not ever making it to the Desktop.. I use Linux as my Desktop and I am very happy and thats all I am really worried about. Do I think it is better? Yes. Do I think it is cheaper? Yes. Do I think it is more stable? Yes? Do I think it is more fun? Yes. But I can't choose what consumers want and it takes much more than just really good code and extreamly pretty GUI looks to sway users ... it takes marketing (yuck!).

RE:Thinking the wrong way
by Timothy R. Butler on Thu 16th Jan 2003 19:25 UTC

Actually, all the system-oriented parts of the GNU/Linux system need to be more friendly, not just X. For example, when you can't easily install new drivers, or when the kernel is breaking compatibility and drivers need to be re-compiled, all that is a pain in the a$$ for the Joe User. It is not just the X/KDE/Gnome that needs refinement, it is all the system-oriented places of the OS.

Its getting there. For example, when I wanted to install a winmodem, all I had to do was grab the Lucent RPM, click on it, type in the root password in Kpackage (which loaded automatically) and enjoy. It would be relatively simple for everyone to do that. ATI's package isn't much harder...

What the Linux desktop needs - My $.02
by Robert Follett on Thu 16th Jan 2003 19:58 UTC

Remember when you had NO desktop with Linux? Remember how excited you were to get X running for the first time? Well, I do! I was holding it's hands for those first few baby steps, and it grew quickly.

The speed with which it has matures is neck-breaking. People all over the planet contribute to it's growth. It is near adulthood now. The problem is that it still has the gawkiness of a teenager, that disproportional akwardness that comes from growing quickly and in spurts. Yet, it simply shouldn't have the charactoristics at this stage in it's growth!

The very thing that has helped the Linux desktop grow toward maturity is the very thing that holds it back; the myriad of people who contribute to it.

I have seen marriages fail because neither party had nothing in commmon in the end, although they did in the beggining. Partnerships fail when the envolved parties don't share the same goals. I see that very thing on the Linux desktop, people going in their own directions. Yes, there are a LOT of people contributing to it's maturity but not with a common set of goals, just individual ones.

I really want to see Linux suceed on the desktop, I really do. Further, I will continue to use it on the desktop because of it's sheer power and stability. However, this fact still remains; the more it matures, the clearer divisions become between people's agendas.

Hm. Eugenia is back. Looks like the party is at full rage again. Here we go!

Mike Hearn: People having to recompile drivers is a two-sided problem, it's only really an issue for closed source drivers (esp the nVidia drivers).

Proprietary/closed-source drivers are a problem not only in Linux, but in any OS. I'm sure lots of people here can't use a peripheral in OS X or XP because nobody wanted to port the code from older releases.

rajan r: I think Linux for the desktop isn't dead. but Mandrake is good as dead. Until they can find a business model that actually works, they are probably better off as a non-profit organization.

Right to the point. The problem is Mandrake itself - what's their business? I doubt even Gael Duval knows.

MP: Second big thing except having better GUI (for Windows users) is to have linux pre-installed on Dells Gateways Compaqs Toshibas desktops and laptops.

I'm sure that if people was forced to install an OS into their computers (e.g. forcing vendors to sell non-OS 'puters) people will see how crappy is installing Windows - BTW, XP was probably the easiest Windows to install, at least for me, but I still think RH 8 install is easier.

Mike Hearn, PLEASE PLEASE PLEASE keep giving us some classes about X architecture (yes I'm serious). I think everybody here needs it.

re:
by Anonymous on Thu 16th Jan 2003 20:39 UTC

<snip> I belive that it was really bad idea to mimic Windows GUI on linux (KDE but also Gnome), because Win user will ask why the hell should he use linux version of windows if Windows (MS) is already here with familiar annoyances including menus and program names and so on and so forth.
I belive that the only way to beat the beast is to make something different and better not as good (FROM DESKTOP POINT OF VIEW) as Windows. With KDE or Gnome linux lost the speed, icons are flashy but for how long are you going to look at them?
</snip>

I couldn't diagree more. People want similiarity to shorten the learning curve, otherwise there'd be more people switching from/to Mac/Windows etc.
The only reason I was willing to try Unix at all years ago was because KDE 1.2 looked so much like Windows 95.

Why ditch X11?
by Timothy R. Butler on Thu 16th Jan 2003 20:44 UTC

Yes, Apple took the Unix operating system and made it accessible to Joe User, but a parallel to Linux it is not. When you install OSX, you don't choose between 11 window managers, 5 office suites, 3,000,000,000 console-based text editors, and 11,000,000,000 command shells. You can't download OSX off the Internet as an ISO for free, and CERTAINLY you cannot download the source along with it. And most important of all, Apple did what it did on hardware in which they hold a monopoly, and they did it WITHOUT X11.

People are always attacking X11, but they can never say why. It seems like it is just "because Apple did it." The thing is, X11 already has tons of apps, is fast on a decently powered system, has great hardware support, and has most of the features that any modern gui should have (including anti-aliasing, easy font config in 4.3, OpenGL, on the fly resolution/refresh changes, etc.)

The biggest complaint people always express is with fonts. But I think fontconfig will eliminate most of these problems. And, once everything is setup, X11's font rendering looks *better* than Windows'.

Yes, manual configuration is cryptic -- so don't manually configure. :-) Anyone who suggests it can't succeed because of configuration issues ought to try to manually configure Windows' GUI. Registry hacking should prove very fun indeed.

The point is, it doesn't matter how cryptic the configuration files are, so long as you have an easy to use config tool. Mandrake has had that for ages, and so do most of the other distros.

KDE needs a bit more work, but already feels like a unified GUI. At the rate it is going, it will probably meet most everyone's needs within two years...

-Tim

KDE's interface
by JK on Thu 16th Jan 2003 20:47 UTC

I agree with the author of that article that KDE has a rather nice GUI. If I could find KDE apps to do everything I want then maybe I would be using Linux now rather than Windows. But at the moment I have to run a mix of different out of place apps each with their own may of doing things. It's often a frustrating mess when I'm working with different apps in Linux, even things like cut/copy/paste don't work in a consistent way. I find going back to Windows is a relief after trying to work in Linux. That isn't just because it's different, it took me about an hour to learn to use Mac OS X comfortably, while I still hate using Linux after weeks of struggling with it.

I have plenty of other issues with Linux (installation and configuration of hardware drivers, RPM installation problems, hardware support, etc.) but most seem to be getting fixed at an impressive rate. However I'm not holding my breath waiting for a consistent GUI and until that's available for Linux it wouldn't be ready for my desktop.

To the Linux on the Desktop Pundits
by Iconoclast on Thu 16th Jan 2003 21:12 UTC

Why does Linux being a good desktop have to be synonymous with joe-retard-user being able to use it? Linux makes a fine, stable, and powerful desktop system just as it is today.

Frankly, I get a little tired of the incessant sissy argument, "Linux will be a desktop contender as soon as they replace X with Windows XP". If you like Windows XP then use it. If XP is so great, stop looking to Linux to provide you with an alternative, and stop trying to make Linux Windows XP. If XP sucks and you want a cheap alternative, then get off your lazy backside and learn something. If all the whiners would spend the same amount of energy learning as they do whinging...

Sorry, I'm calm now.

Re: origins of NT
by DrP on Thu 16th Jan 2003 21:14 UTC

>>NT has its origins in OS/2.

>This is false. Win9x has its origins to OS/2.
>NT has its origins on VMS. It was in fact designed by the >same person who did VMS.

Well, VMS influence may have been in there somewhere (like from DOS days), but NT really did come from OS/2. Remember when IBM and Microsoft had their "divorce" over Windows? IBM continued to develop OS/2 2.0. By mutual agreement, MS took what was called "OS/2 3.0" which became the first NT. As one other poster noted, that is why you could run OS/2 apps (Character based) in NT. One thing NT admins are taught to do for security is strip out the OS/2 compatibility layer.

Win9x bears no resemblence to OS/2. Win9x is just a semi-32bit shell on 16-bit DOS. OS/2 is certainly not that and never was.

The reality is...
by Ralf. on Thu 16th Jan 2003 21:39 UTC

if you like to hear it or not:

Linux on the non geek end-user desktop will not happen. And I think that is not the intention of Linux.

And - if you like to hear it or not - the only Company that managed to deliver a Unix with a end-user frendly desktop and system-management is Apple.

Sorry,

Ralf.

Easy to use UNIX
by JK on Thu 16th Jan 2003 22:06 UTC

I really hope other Linux users are embarrassed by Iconoclast's ranting, people like him are the reason Linux fans have such a reputation as ignorant elitist brats.

Ralf: Apple were hardly the first to make an easy to use UNIX, I think Apple would still be working on their new OS if they hadn't had NeXTSTEP as a basis for it. SGI IRIX was rather nice too, and BeOS was UNIX like yet easy to use. I don't think there's anything about UNIX that makes an easy to use desktop hard to create. IMO the main problem with Linux is the lack of standards and it's immaturity as an end user desktop.

The Irony of it all
by Sikosis on Thu 16th Jan 2003 22:07 UTC

Wow and I just installed Mandrake on 6 of our PCs at work to replace Windows XP Pro ...

Xfree86 *sucks* and here's why:
by riggits on Thu 16th Jan 2003 22:19 UTC

A) memory hog memory hog memory hog!! take a look at all your processes running under whatever distro you care to name and tell me that Xfree is not bloated. Case in point.. however, just for laughs, let's continue.

B) responsiveness (perceived) is incredibly slow. Why? Well, that's kinda obvious... Xfree is a single-threaded bloated app that then further subdivides its timeslice to update all those GUI apps you run. No latency is assured, which means that high latency is the norm. Windows 2000 avoids this mess, as did BeOS, as does Mac OSX. And to the retard that thinks that Apple supporting Xfree under Mac OSX is validation of his precious Xfree: "support" means you can run it, not that it runs well. Think about it. In fact it runs like the proverbial sh*t from your rear end, only not so smelly.
As an aside, please notice that Apple avoided Xfree in favour of a far superior architecture. Good choice, Apple.

C) Cruft. Xfree is spaghetti. Who understands it, other than those dinosaurs who have coded it all their lives? Just take a good look at Xfree (how many millions of lines of code now? I have no idea) and if you can decipher it then you da man. The devels don't shy from open admission that there's way too much cruft in the code.

If the designers of X-Windows built cars, there would be no fewer than five steering wheels hidden about the cockpit, none of which followed the same principles -- but you'd be able to shift gears with your car stereo. Useful feature, that.
- Marus J. Ranum, Digital Equipment Corporation


Check out http://catalog.com/hopkins/unix-haters/x-windows/disaster.html for more. Don't let the "unix-haters" bit throw you off - this isn't a troll.

nope
by fredrik on Thu 16th Jan 2003 22:57 UTC

>>NT has its origins in OS/2.

>This is false. Win9x has its origins to OS/2.
>NT has its origins on VMS. It was in fact designed by the
>same person who did VMS.

Indeed NT was designed by some of the ppl who also designed VMS, but it contains no VMS code. However, MS (who owned part of the OS/2 code) used their part of what would have been OS/2 NT to have as a base for NT, also NTFS is based upon HPFS IIRC. You could still run OS/2 1.x apps on NT 4, not sure about win2K and XP though.
http://www.winntmag.com/Articles/Index.cfm?IssueID=97&ArticleID=449...
win9x contains no OS/2 code whatsoever!

riggits, you are trolling after all
by Datschge on Thu 16th Jan 2003 23:22 UTC

I assume you never used Linux or any graphical Unix before. Get Knoppix at eg. http://www.knoppix.com and tell me X Window is slow. And as for your cruft comment, you did the best to prove that all your proofs are crufts themselves. Quoting a 9 years old book while we are debating about up-to-date computer technologies? Do you have any pride?

-----
The Unix Haters Handbook is published by IDG Books Worldwide, Inc. 155 Bovet Road, Suite 310, San Mateo, CA 94402. Copyright (C) 1994 (!!!) by IDG Books Worldwide.
http://catalog.com/hopkins/unix-haters/handbook.html

I actually own that book. It's out of print and actually quite valuable. If you don't believe me, check it out on Amazon.com. While I have read it many-a-times and think it's quite funny, nearlly all the problems described in there are problems with the traditional UNIX's. Keep in mind this book was published in 1994, and most of the material in it is from the late 80's. Things have changed since then. The new UNIX's like FreeBSD and Linux don't operate that way anymore, and XFree86 just had a total rewrite in version 4.

The point is that this book is in no way meant to be a technical argument against X11, it's just a book of jokes, and shouldn't be taken any other way.

Further, X11 is only a specification, and there are MANY implimentations. XFree86 is probably one of the most modern ones. To clearify, XFree86 uses sockets for local ICP for compatiblity reasons. Not all systems have implimeted System-V ICP. Linux has message queues, but they are slower than sockets currently (which should not be). Nevertheless, XFree86 already has it's own shared memory extension. XFree86 is certianly not slow. I'm tired of hearing why XFree86 should be replaced from people that have never written a line of code for it and know nothing about its workings. I'm not saying that there is nothing that can improve, but I am saying that there is nothing wrong with it that can't be fixed.

RE JK
by Iconoclast on Fri 17th Jan 2003 00:02 UTC

I really hope other Linux users are embarrassed by Iconoclast's ranting, people like him are the reason Linux fans have such a reputation as ignorant elitist brats.

"Let us pity the ignorant Linux user who is too simple to understand the thoughts and workings of us superior Windows users." Indeed.

Instead of calling me names, why don't you answer my question. It is this kind of impotent response that doesn't do anybody any good. Now if you want to answer my question as to why Linux can't be ready for the desktop until Linux == Windows, and why Linux can't be a good desktop until every joe user is born with a Linux-using instinct, eliminating the need to actually learn how to READ, then I'll listen; otherwise, it is all the same noise.

Ease of use
by JK on Fri 17th Jan 2003 01:53 UTC

Iconoclast: When has anyone said that Linux needs to turn into Windows? You're making a meaningless strawman argument.

I certainly never said Linux wasn't a good desktop for some people, but that doesn't mean that it doesn't have significant ease of use problems. fixing Linux ease of use problems and turning it into Windows are two very different things.

At the moment Linux simply requires much more work from the user than Windows or Mac OS IMO. Sorry, but most poeple have better things to do with their lives than read man pages and FAQs to get things working. If Linux is going to become a mainstream desktop OS then it's problems need to be fixed, sneering at "joe user" for not being a big enough geek isn't going to change that. If you don't care about Linux being sucessful on the "average" user's desktop, then why are you posting in this thread?

I really can't understand the BS that some Linux fans spout about Linux being "dumbed down" or "turned into Windows". How do better graphical tools and easier configuration damage the power of Linux? It just seems like a bunch of sad geeks who want to keep Linux as their elitist toy, so they can look down on users of other OSes and feel smug and superior.

nice try Xfree
by riggits on Fri 17th Jan 2003 04:19 UTC

I've used about ten distros and they all have terrible performance problems on my old P3-600 laptop, and I suspect that the lack of RAM is not a problem.. I've used Xkill a few bazillion times, and who hasn't? Linux may be stable, but all the crap on top of it sure isn't.

When will automount work? Probably never.. it's been languishing with no discernable progress for years now.

Will USB support surpass Windows 95 OSR2? Maybe next major kernel number.. maybe never. Drivers will also take their time in coming, through no fault of the developers, but the truth is that if my webcam or mouse (Microsoft Intellimouse Explorer, for the curious) won't autodetect on hot-plug then the OS ain't worth a damn to me.
Of course, let's assume that automount gets fixed, USB support improves (which it inevitably will.. but will it still matter?) and maybe even - just to be nice - let's say that Linux users get treated to a proper installation utility similar to the plethora of options available for the past five years on Win32.
Will it matter? I can't run my programs, like SoundForge, ACID Pro, Reason, Vegas Video, etc. ad nauseum.
That alone isn't so bad.. after all, there are plenty of great GNU alternatives, right? say the zealots
Wrong. Lucky Linux users get the GIMP, with about 10% of the functionality that I personally require on a monthly basis from Photoshop. They get ... what? I can't find a single sequencer worth using from the GNU stable.. there are a few copout wannabes for Win32 (eg. Psycle, Buzz, whatever) that are free and won't do the job. That doesn't cut it.
Basically what I'm saying is "show me the money!"
WINE blows, as anybody who has used it can (but probably won't) attest, and even Office isn't stable under WINE. Games are slower too. So don't tell me to use WINE for all my Windows needs... it ain't workin' yet.
Xfree is slow. How slow? Too slow to be useful. Enlightenment?? Don't jest! The turkey who suggested that one has obviously never tried it. Bloated alpha-level software is no replacement for a proper GUI.
Other window managers are equally useless for my purposes, and the lack of programs here doesn't even enter into the equation yet - KDE is a bloated huge piggy-wiggy, and anybody who claims different has probably only read the propaganda. GNOME is somewhat better but equally irrelevant and it's easy to see why: following the leader thirty steps behind isn't good enough anymore. Even if your product is free. If it were, Linux would have soared by now.
All the same, I wish Linus and the boys good luck... it's a neat kernel he's got.

nice try riggits
by Datschge on Fri 17th Jan 2003 07:39 UTC

I'm working with KDE3/Linux on an AMD K6-2 300mhz on a regulary base and it's by far the most stable computer and smoothest among the computers I'm forced to use (mostly Windows and a Mac at work). Similar cases are there for many, whether it be KDE, Gnome atop of whatever distribution, possible even one of the *BSDs which is still said not being ready for the desktop even after Apple is putting their whole OS line atop of it. Don't keep telling us what's bad if you want to keep using your Windows anyway. Just posting for the sake of batching other people is silly and a waste of time.

Anyway the only showstopper for Linux on desktops imo is binary compatibility. Nothing more than that is needed to get more commercial support. But I doubt we'll see binary compatibility anytime soon as long as most distributions see themselves as all included system which don't need to care about binary compatibility. http://www.smh.com.au/articles/2003/01/15/1042520666517.html

And btw, Linux is the kernel, nothing more.

To all you X-Haters...
by Jeff Moody on Fri 17th Jan 2003 20:14 UTC

So you think that XFree is inherently flawed, slow, clunky and obsolete?

Then why not start working on the alternative?
http://www.picogui.org/ - PicoGui is a project started originally for the Sharp Zaurus, but has been shown to run on the great majority of device architectures that Linux itself runs on.
If folk were to take the time and make efforts in (1) maturing the PicoGUI infrastructure and (2) porting toolkits from X to PicoGUI, then you've got your damn alternative.

Quit complaining.
What is the old saying? "Actions speak louder than words"?

REAL FLAW
by JEFFERY on Sun 19th Jan 2003 08:09 UTC

Mandrake's real flaw on the desktop? It couldn't log onto yahoo and play pool. Not it's fault? Minor problem? This is the type of thing that users want an operating system to do "out of the box", and there's no reason why it shouldn't. Attention to detail is everthing. Linux vendors don't get this.