Actually, not anymore. Mandrake 7.0+ has branched off RedHat, and 7.1 is QUITE a bit different from Redhat. (It has ReiserFS, it has XFree 4.0, and the other packages are up to date.)
People who think that/. has every right to post messages before people have a time to mirror piss me off. There is a thing in this country called "responsible journalism." Since the journalistic community these days is in such a rut, I'm not surprised that most of you haven't heard about it. In traditional media, responsible journalism takes the form of corroberating your evidence to make absolutely sure that you're giving the correct news. Until recently, all respectable establishments did this. Those that didn't are called tabloids. Because of their rush to get the dirt, tabloids often post too early about a story. If it turns out to be true, then the tabloids will often have the news before any of the papers, but if it is false, then it ultimately hurts those that the story is about. In the new electronic media, journalistic responsibility has to extend to thinking about the ability of servers to take an increased hit before they are ready. It is analagous to allowing a defendant to prepare his case before the prosecution starts. In the good old days, the media had responsibility. They would rarely post a news story about, say, a president before he had made a public statement about the situation. The rumors and such were left to the tabloids. This situation is similar. You don't attack somebody (the/. effect is certainly an attack, though unintentional) before you give them a chance to prepare. It is just not decent./. posting new files before the mirrors have a chance to gear up is not just impolite (or I'm I just a dinosaur and curtousy is passe?) it is also detrimental to the community, which wants to be able to download the software.
Reiser FS looks seriously cool. I tried it out a little on Suse 6.4 and it flies. However, there are a few things about it that piss the hell out of me. First, it isn't well integrated. I don't know why, but for some odd reason, LILO can't boot a reiser filesystem, so you have to make another partition for/boot. Its a pain, especially if you've got 3 OSs like me, and hit the 4 partition limit way way to often. I have to wonder, though, what the hold-up is. I'm not trying to troll here, but what's taking them so damn long? They've been working on it since.99 and its still not done? BeOS already has a from scratch fast, fully journaled FS that's been in there since at least R3. Also, it already has the database capabilities that won't make it into Resier for a while now. Why's the project taking so long? It's not like they're perfecting it, ResierFS still is less stable and loses data more often than BFS. Still, for Linux speed-mongers ( he he, oxymoron) Reiser is a must have.
Actually, a fully journaled filesystem need not be slow. Take a look at BFS. It is freaking fast, and fully journaled. (Hell, I've given up shutting down. Be should just remove the shutdown menu.)
Oh yea, Linux is just so hideosly great. It can't match BeOS in response time (their midi kit now has microsecond response and only RTLinux can beat its audio latency.) It still doesn't have a GUI that is as elegant and inuitive as Be's, it is just now gettings direct access to graphics hardware (not just a simple frame buffer) and most relevantly to this article, it still doesn't have a journaling filesystem that is as fast as Be's, as stable as Be's, and has a built in database engine. Seriously, Linux is great and all, but before you spout bullshit, remember that there are a lot of areas (SPEED!) where it is not #1.
By opening up the playstation hardware, Sony has made a great "in" for itself. In the console market, nobody ever makes money on the hardware, they make money from software. On the PSX, Sony gets $7 for each CD sold. It only makes sense that if more people have conoles, Sony will make more money. As for the X-Box, I'm not to impressed with it. It looks like it will be a regular PC CPU paried with a regular PC graphics card, running a version of NT. Is it just me, or does this scream "innefficient." Face it, people are never going to want to edit documents or photos or use a console for normal PC tasks. The console will always be a reciever for content, wether it is from the web, a game disc, or from a DVD movie. Sure people might edit a few emails, but this whole integration of one stop photo editing, document writing, etc, is unlikey to happen. After all these years, your toaster is still seperate from your microwave is it not? Most people have both a toaster and a microwave, and I'm pretty sure that in the future, people will have both a set top box and a computer. By making the XBox so general purpose, MS does two things. One, it brings itself to the game late because of the extra development time. Only one company has ever been able to pull itself to the front after coming in late and that was Nintendo (then the 800lb gorilla of the market) with its SNES. Now MS is not the 800 lb gorilla of the console market. Sony is. By coming to the game late, they severely hurt themselves, because for a console, all that matters is quality and amount of software. Nintendo found this out the hard way with N64, and MS will likely get a beating from Sony because of this. Also, MS brings too much PC crap to the table. Face it, if NT isn't stable enough for your mission-critical server, it isn't stable enough for your console. The 300 days between crashes for a Linux box is megear compared to the 3 years my PS has gone without crashing, while being run at 100% by funky custom ASM code no less. The PC design also brings a lot of overhead. The PSX2 architecture is much more integrated than that of the XBox. Integration means a lot. Take a look at the N64. This thing has better graphics than a Pentium 200 with Vodoo and does full screen antialiasing and graphics features that PCs only dreamed about until the TNT era. It does all this with only a 97.5 MHz R4300i and a 66MHz graphics proc. If you take into account the efficiancy of the architecture, you'll notice that the XBox doesn't look so hot anymore. I'm willing to prognosticate that in the end, it will look like this. Sony comes out with the PSX2, makes a ton of money, gets a huge installed userbase, then a year and a half later, the XBox comes out. It is maybe about 30% faster, but doesn't have the huge gaming library, and is a year and a half behind Sony in building a userbase. In addition, people get frustrated by having to deal with PC-type technology (ie. crashing) and find another reason not to by the XBox.
UNIX is NOT user friendly if you don't do the things that UNIX doesn't like. Windows is not friendly if you don't do the things that Windows doesn't like. If rm -rf * is your style, the UNIX is your game. If right-click->Delete is your style, then BeOS is your bag. Windows is only your bag if you need to run a program. If you're running a server, good luck if you try to do it in windows. If your running a desktop, god help you if you try to run it on UNIX. I am quite experianced with Linux, but it still amazes me how much crap I have to put up with to do simple desktop things like install a new 3D card driver. (I'm not kidding by the way, it is ridiculously easy in Windows. More so in BeOS where the system automatically installs drivers.) Of course I've dabbled in networking, and am amazed at how easy UNIX makes it.
Really, you use an old mac to do website stuff? Increadible. I just buy new hardware. I am performance obsessed too, but I wasn' talking about a pixel level editing program. Neither was I talking about operating system-level stuff like GNOME. (Wether Linux people like it or not, for all intents and purposes, the DE is part of the OS.) I was talking about the 3D renderers and the Office apps, and the web browsers, and the photo editors and the video editors. The meaty stuff that people use every day. Sure there are those who have powerful hardware, but use older programs for the sheer speed, but unhappily, they are a rare case. My point was that developers know what kind of CPUs their stuff will run on (at least in the mainstream, I don't know about UNIX though) because people tend to run modern programs on new hardware. I am sure that the developer of that plain X app had no idea how powerful a PIII would be or how to optimize for it. But that's okay, because 99% of the time, nobody will use that old X app on a PIII. The point is that the compiler doesn't really have to worry about optimizing for different architectures from the future (in the same instruction set) because the program will rarely be used on those architectures. The post I was responding to thought that dynamic compilation would win out, because when the K8 or K9 comes out, software written for the older CPUs would be optimized better for these new CPUs. A static compiler can't do this, because it can only optimize for CPUs that are known at the time. I was saying that it doesn't matter, because people probably won't be using that app when the K8 or K9 comes out. Even then, the software won't need optimization, because it will have been writtent to run well on an older CPU.
I don't think you realize that you aren't talking about dynamic optimization, you're talking about optimizing compliers in general. A dynamically optimizing complier has a few edges over a statically optimizing complier, but in general, the static compiler wins. Think about the differences in hardware archicture. Aside from something like an OS, a developer already knows what CPUs their program will be used on. If I make a performance critical program now, say a 3D renderer, I know that it will be used on PIIIs, PIIs, Athlons, Willamette, and Maybe K8. That's it. People will not be still using my program on K9s and 10s. Sure some people use older programs, but think about it. In programs where performance really matters, do you use anything older than two years?
The poor showing of Visual C++ was very surprising, given that Visual is held to be one of the fastest x86 compilers. The fastest x86 compiler at the moment is Intel's own x86 compiler and it would have been interesting to see how it would fare on the benchmark. I have done performance tests between Visual and Intel's on graphics apps on a PII, and Visual's is rarely more than 5-10% behind Intel's. A future extension to this could be to test this code with Intel's compiler since it is downloadable on the web (full but time limited) and easily pluggable into Visual. I suspect that the weakness of the Visual complier may have something to do with Microsoft not optimizing as much of a non intel CPU and the recent K7 optimizations gcc has gotten.
How 'bout the other way around? Think about it this way. There are no serious Linux 3D applications that can use the 3D accelerated drivers, (since most static link Mesa) so the only major 3D thing it can run at the moment are games! I'm sure this will change as 3D apps ar ported to Linux, but for now, Linux has nothing that wants to make me trade in MAX. Also, he was reffering to those running XFree 4.0 in general, not just the 3D stuff. Trust me, if you are a desktop user, you need XFree 4.0. I have not yet seen a crash and it is much faster than 3.3.x. In fact it restored my faith in Linux as an OS when I saw that finally XFree86 could blit stuff to the screen as fast as DirectDraw (windowed) could.
This is awesome, except not really. It's great that a company is offering Linux support, but an announcement like this in the Windows world would be met with laughs. SIS has never made even a mediocre graphics chip. All their chips have rated from Man this sucks to Man this sucks as much as Windows 2000! It really shows how easily exited the alternate OS community is about any support whatsoever.
KDevelop alone is worth the price of admission (which of course is $0.)
I tried it last winter, and found it utterly useless for creating anything but KDE apps. I'm trying to write ISO C++ libraries, and Kdevelop (at the time, at least) didn't even have a library project template! Maybe I'll try it again sometime, but I was seriously disapointed with what I saw. >>>>>> Huh. If I'd tried GNOME at that version level, I'd have run away screaming. Try anything 1.1 and above. It is rock solid, and now (in 1.2) has integration with GNOME development files. It is a serious competitor to VisualC++ if they can only get rid of that annoying "I" beam, and get IntelliSense type stuff in there. I develop entirely with classes and it is a bitch remembering function headers. This is not meant as a slam towards the Kdevelop people, for creating a KDE app it would be great. It just didn't do what I needed it to do. >>>> Try using an app when it is release-quality. I don't judge Linux on my impressions of Slackware 3.5, and you shouldn't either.
BTW, why did you mention Kdevelop, when you claim you are talking about the general desktop market, or as you put elsewhere: "Linux grass-roots zealots weren't invited. >>> I was talking about the general desktop market. Many programmers are part of this too. Real grass-roots zealots use VIM:)
Correct. However you conveniently ignore the fact that the reverse can also occur. Will it? I don't know. But it's about as likely as your scenario is. >>>>>> Not really. There is a thing called a landslide effect. Once the tide significantly begins to turn in one direction, it is unlikely that anything short of a revolution will overturn it. KDE is significantly ahead in the apps and developers game, and it doesn't look like GNOME has anything on the horizon that will turn the tide. (Evolution will only take you so far.)
Have you by any chance used Gnome 1.2? Didn't think so. Go use it. Then come back and say that again. Or don't, if you suddenly find that statement you made holds no water whatsoever. >>>>>>>> I have, and my statement still holds water. GNOME 1.2 is still slightly slower than KDE. True, it is better in this release, but here is a good test. Take an app that has a lot of views (like ksysv or whatever the equivilant GNOME app is) and resize it. Watch the redraw. See which one redraws faster. GNOME 1.2 is about on par with KDE 1.x in terms of speed, but KDE 2.0 is actually faster than KDE 1.x on my machine.
You're some sort of psychic, I take it? It seems to me as though Gnome and KDE are making the exact same types of changes along their paths to 2.0. New file managers, component architectures, themeability on KDE's side, etc. Gnome's added a 1.2 step, but what's the harm in that? >>>>>>>>>>> Yes, I'm phsycic. My point is that KDE 2.0 has less memory usage than GNOME 1.x. Since KDE 2.0 already has all the new stuff in place, it is doubtful that it will grow too much in the future. GNOME 1.2, however still doesn't have all the features in place, and can do nothing but grow in 2.0. Thus, KDE 2.0 will take significantly less memory than GNOME 2.0.
And you speak from... how much experience? As I thought, none. Go use it, then come back when you can make arguments and back them up. If you can't back up your arguments, you're just flaming. >>>>>>>>>>>>>>> I don't speak from firsthand experiance with either environment, true, but I can tell you this. Bonobo is similar to KDE's original object architecture, which was based on MICO. They ditched it in favor of the much simpler and faster DCOP arch. I doubt there is much the GNOME guys can do to make an inherently slow architecture (even the Berlin guys admit this) fast.
Ah, and thus we get to the whole crux of your argument, and the real reasn you're posting: "KDE is better because I'm a Corelite." Let's see your examples. I hve no doubt they probably exist, but where are they? You're certainly not too cooperative in pointing them out. Why not? >>>>>>>> My first Distro was Slack. My second distro was RedHat 5.1. My current distro is Suse 6.4. I'm running KDE 2.0 beta 1, on kernel 2.3.99pre8 with the beta nVidia OpenGL drivers hooked up to XFree86 4.0 all running on ReiserFS. I am certainly not a Corelite. Even an advanced user appreciates cohesion in their work environment. Plus, it has nothing to do with advanced users anyway. My discaimer at the top got cut of for some reason, but I wasn't talking about the "core" linux crowd. I was talking about the new Linux users (ie. the business) market, those who couldn't care less about GPL vs. QPL and only wanted a fast, stable, COHESIVE desktop. If you use Linux because you love it, then this post doesn't concern you.
According to whom? KDE has StarOffice, but other than that every KDE app I've seen has a Gnome analog. For the most part, the revere is also true. >>>>>>>>>>>> KDE has nothing to do with star office. KDE has KOffice, which by itself gives it an enormous lead over GNOME and whatever it's office suite is. (A mish mash of Abiword, Gnumeric, Evolution and a few others.) Aside Evolution, KOffice looks like it will blow the others away. Then, of course, there is KDevelop which again is a "killer app." Both platforms may have similar apps, but in general the KDE ones are higher quality/more useful.
Again, you conveniently ignore the fact that the people could just as well choose Gnome as KDE. >>>>>> My whole post is about why people will choose KDE over GNOME! This point only accentuates the fact that people (most) will choose ONE, not live with both.
Which is...? Your answer is conspicuously absent. As if you realized that since you have near-zero experience with Gnome, particularly recent versions, you just might be wrong. Use them, if only for a little while, so you can make a real comparison in that regard. >>>>>>>>>> KDE has more and better apps. I've been over this already. KDevelop, KOffice, and the great set of apps bundled with KDE are much better than their GNOME counterparts. Do I even have to debate the merits of KDeveloper of gIDE? I have a lot of experiance with GNOME in fact. I've been using Helix GNOME on RedHat ever since it came it. I had used KDE for a month before that, but liked GNOME's asthetic sense much better. Near the end of my RedHat cycle, I tried KDE 2.0, and switched to it and Suse 6.4 a week ago. (And yes, I do have both Helix GNOME and KDE installed.)
And please don't forget that even if you are correct, that means little. Remember, people stuck with Windows, which hardly has the "best" apps (the single possible exception being Excel, and both Linux DE's have spreadsheets which are coming along quite nicely in that regard). >>>>>>>>>>> Are you kidding? People stuck to Windows because it has the best apps. Where it doesn't (ie the high power server market) people don't use Windows! Are you telling me that Gimp beats photoshop, or anything on Linux can beat MS Word, 3D Studio MAX, Maya, SoundForge, Cakewalk, Truespace, IE (it's true, Netscape sucks), Visual Studio (if you're into the IDE thing), and the limitless games available for the platform? These are best of breed apps and are ONLY available for Windows. People don't use Windows because MS forces them to. People use it because it has great apps. People like me reboot into it because they have to use 3D studio or photoshop. (Because they rock!)
For better or for worse, Gnome is just as Windowsy as KDE. And both are getting even more Windowsy as the versions progress. It's quite sad, really. You'd think the Open-Source community could do better, interface-wise. It's been done. NeXT had a better GUI which was almost completely original. So did MacOS. Win9x and OSX are fusions of these two; Win9x took the worst of both and added some decidedly anticompetitive elements. It'll be interesting to see if OSX has it done right (I admit I don't like some of what I see, but I'll reserve judgement until I've actually used the thing). >>>>>>>> You miss my point. I wa critisizing people who think KDE is too Windowsy. I neglected to mention GNOME, but sure, why not. People could really care less about the quality of the interface. People like something that is familier, no matter how hard it is to use. (Of course, they won't bother to familiarize themselves with something that is harder than what they are using now, but also won't switch unless the competitior is revolutionarily easy. People are lazy.)
Show me your "growing support from the business community." I don't see it. I saw StarOffice getting KDE integration back when it was still proprietary (versus the pseudo-proprietary Community-Source liense it's under now). Oh, and Corel bundles it with their system. That's all I ever saw. If there's more I'd be glad to hear about it, but I simply don't see it anywhere. >>>>>>>>>>>>>>> Oh, and Corel bundles it? Like that is a little thing? Corel is the only respectable (traditionally) company doing Linux, and you act like it is a small deal? Also, must Linux distro's have much better support for KDE than Linux. Go to Mandrakes website. Compare the number of GNOME screenshots to KDE screen shots.
Give an example of a "big, memory-hogging" feature Gnome has that doesn't make it better. >>>>>>> Bonobo. It makes it better, but not enough to warrent the weight.
Depends. Everyone knows that Gnome 1.0 was released too early; I can't dispute that claim. It was underfeatured and unstable. The stability got fixed quite some time ago. 1.2 adds more of the features. Not all of them yet; a few are still in development (Bonobo and Nautilus, most notably). >>>>>>>>>>>> It took GNOME until 1.2 to become as stable and nearly as fast as KDE was almost a year ago. What does that tell you?
Speed... well, I'm not so sure. A lot of that seems to have to do with Imlib. Everything that I've seen that replaced it wuth gdk-pixbuf seemed to get a big kick in the pants, speed-wise. >>>>>>>>>>>>>> Great. 2.0 will be faster. But so will KDE 2.0. At everystep, KDE seems one step faster than GNOME.
Of course, you can't expect phenomenal speed when you're drawing your entire GUI with pixmaps. That's a large reason of why Gnome feels so slow for a lot of people. Switch out Sawfish for Window Maker and the Pixmap themes for GTKStep and you've suddenly got a much faster desktop. I like Sawfish myself, and used it for several months, but as of this moment it just doesn't seem to be able to touch Window Maker for speed and stability, though Window Maker's icons do get in the way on a Gnome desktop. Things may have changed with the developmental gdk-pixbuf versions of Sawfish; I intend to try it out as soon as I manage to compile the thing. >>>>>>>>>>>>> KDE 2.0 can draw the entire desktop with pixmaps much faster than I've ever seen GNOME do it. Plus, my themes to use pixmaps, so very little is being drawn with pixmaps.
Either way, this is what's wrong with the state of discussion on Slashdot. People don't debate anymore; they just flame. They'll spout drivel like this post's parent, which say a lot but prove nothing, and thus are no better than "Gnome sucks! No, KDE sux0rz! No way, Gnome 0wnZ KDE! No, KDE is ph4r m0r3 k-r4d @nd l33t! No, you must PH3AR GNOME!" What happened to people not jut making their point, but backing it up? OK, so Gnome sucks NINJA ass; why? OK, so maybe KDE sucks big fat naked petrified donkey dick; why? The point is, unless we get into real discussion, and point out the relative strengths and weaknesses of the two systems, neither one is going to truly improve. They'll just get stuck in an endless feature war, neither one getting the features users really want. They'll just try to one-up the other, and the end result won't be any better than Win2K. >>>>>>>>>>>>>>. I just pointed out a dozen weaknesses of GNOME. I'm not saying KDE is flawless, but GNOME is still much worse. I'm sorry if that's not what you wanted to hear, but for the mainstream market, (my disclaimer got cut dammit), KDE is still better and getting better faster than GNOME is!
My account info got lost so I m posting as an ac (e-mail mike@redtux.uklinux.net) >>>> Sure
Points in no particular order 1. KDE is faster? In my experience KDE has never been faste. I recently installed mandrake 7 on my work box with kde 1.1.2 and october gnome. KDE is still far, far , far slower than gnome to to do anything useful. I don,t care how fast it redraws the screen, ftp transfers in terms of bytes in a term I do care about. >>>>>>>> Doing the French add an 'e' to every word thing? Are you a desktop user or a server admin? Of course screen redraws matter. Whats the point of using Linux if your GUI is going to be slower than Windows? And what does the DE have to do with FTP transfer? If anything, it would be faster under KDE due to fact that the ftp program would have more memory left. 2. Gnome unstable yeah!! nuff said >>>>> Face it. GNOME stability sucked until 1.0.50 came out, and even then it was crappy. The first stable GNOME (by Linux standards) was October GNOME running a WM other than enlightenment.
3. Better integration of KDE apps. I use Linux - I do nort want a monolithic desktop. >>>>> Did you read my disclaimer? Linux grass-roots zealots weren't invited. I was talking about KDE and GNOME in terms of a viable desktop environment, not something that fuels some nerd's desire to be 'leet just because his apps don't match.
BTW I stopped using KDE at 1.1.1 because it was so slow and buggy and went back to fvwm and then started using gnome at 1.0 >>>>>>> The sheer fact that you'd fvwm shows that you aren't a desktop user, and the fact that you think GNOME 1.0 was anything but alpha software (again by Linux standards) shows that you're on crack.
One of my points was that it was all about the apps, and KDE has the good ones. Though the Gimp is probably irreplacable, KDE has the majority of the good apps. KDevelop alone is worth the price of admission (which of course is $0.) In addition KOffice is shaping up startingly quickly, and though the spreadsheet may never be as good as gnumeric (doubtfull, as the KDE guys seem to know their apps pretty well) the rest of KOffice will be a major factor. Also, KDE has a lot more developer support than GNOME does at the moment, and in general the KDE apps are far more mature than their GNOME counterparts. (Aside from a few exeptions like Gimp and Gnumeric.)
Every time I see another release of GNOME, I shudder to think of what could have been. At this point, it is obvious to me that GNOME will lose the war with KDE in the desktop environment department. Now don't start on me about how free software cannot lose a war, or about how the two are not competing. The one that loses is the one that has the fewest users and a declining userbase. Since both DEs have their own API, a DE can lose the battle if nobody develops for that API. If most people start using KDE, then few people will bother to develop for GTK/GNOME, thus GNOME will have lost the battle. You can say that GNOME is a project and will continue wether or not it has support, and you'd be right. A project can still be alive and still have some users and still have lost. At some point, it just become irrelevant. Take a look at OpenStep. Sure it may oneday rise up and grap the hearts and minds of Linux users, but it is more likely that it will continue to putter along, maybe contributing some valuable technology to one of the dominant DEs in the proccess. Even if the developers do not want it to be that way, GNOME and KDE are at war. They are at war for the deskspace of the user and the coding time of the developer. Up until recently, the only reason I got anywhere near GNOME is because it had the nicest user environment available in Linux. It was asthetically pleasing, and generally nicer to work with than KDE. Of course, in KDE 2.0, KDE took GNOME's asthetics and usability and made it its own. Now that both are on the same level, interface-wise, KDEs strenghts can come out and allow it to pull ahead of GNOME. In a succinct list, these strenghts are, 1) Speed and resource usage. KDE 2.0 is by far faster than any version of GNOME, and takes a significant amount less memory. This disparity will only widen in the future, since the general GNOME infrastructure will grow quite a bit in 2.0, while KDE's probably won't grow much beyond 2.0 since the basic functionality is in place. Couple this with the fact that DCOP/KParts is far superior to Bonobo in terms of speed and resource usage (though maybe not as flexible in a distrubuted environment) KDE bowls GNOME over in the speed/resource usage deparment. This is significant becuase people don't leave Windows to arrive at another bloated OS. 2) KDE still has a much tighter integration between apps than GNOME does. It only takes a cursory look at Corel Linux to show what a little ingenuity and KDE can do to make Linux almost as friendly as Windows for the desktop user. 3) KDE has far better apps and much more developer support. Sure the KDE libraries can be loaded, but with the increased bloat in both 2.0 level libraries, people will increasingly wish to not have to load the libraries twice. As such, they will stick to the DE that has the most/best apps. 4) KDE is more familer. Diehard Linuxites call it being too Windowy, but it is a big strength for the rest of us who grew up on the start menu and Explorer. It is a moot point which one is theoretically easier/more intuitive. If that mattered, NeXT would still be around (or AfterStep would be dominating in the DE department.) Whatever is easier is whatever people are used to. MS might have the clout to force a new style of desktop on people (a la Windows 95) but Linux does not. These factors, coupled with increasing support for KDE from the business community will allow KDE to become the dominant Linux desktop. Back to my thoughts about GNOME, it saddens me to think of what GNOME could have been. It has a great set of people working on it, and the developers have a great asthetic sense. However they seem to think that putting the biggest, most memory hogging features into the DE will automatically make it better. Software size and quality seems to not be a very top priority, and software speed seems to be an even lower priority. Features seems to head the list, and everyone knows what that mentality gets you... it gets you Windows 2000.
It could, but PCMCIA support is quite a bit easier. Also, the kernel still has to have some sembelance of what PCMCIA is (in this case a bus) in order to load the module. Sure DirectX could be loaded as a module, but it would still need a great deal of access to hardware without going through the kernel, and would still have to be an integral part of the kernel. It might be loaded from a module, but the kernel would still have to include new code to allow the module to take over the system. (What do you think DX is for? Total control baby;)
DirectX would have a hard time being ported into Linux. Like NT, Linux has a significant admistration layer, and it wouldn't meld to well with the rest of the OS. Unlike OpenGL, which has a higher level interface to the OS, DirectX has to be down there in the core. Unless the Linux gods officially bless a DirectX port and integrate it deep into the kernel, it will be hard to get a good, fast DirectX port going without forking the kenrel.
Now for my obligatory BeOS spiel. Wine is not only helping LInux and other Unicies, but alternative OSs in general. In particular, Wine has already been compiled on BeOS, and things are moving at a fair clip to get the thing working. Wine is really a godsend to some of these alternative alternative OSs because it will allow them (when it's done) to have a body of functional software that allows users to complete some important tasks. Unlike Linux, these OSs are missing a few critical apps, and what better way is there to quickly fill this gap than get good emulation (I know it's not emulation per se, but for lack of a better word) going so Windows apps can be used. If you have any coding talent, support the BeWine project at bewine.loungenet.org. It seems that it is pretty low profile and not enough people know about it.
I wasn't talking about PC OEMs. I was talking about the lower end workstation OEMs like Compaq, SGI, IBM, and HP that have high end UNIX machines, but had adopted NT for its lower end, especially workstation lines. Moving to Linux for these machines gives these companies much more flexibility and power than if they had used NT. I know that sgi hated using NT for its machines, and I'm sure Compaq, IBM, and HP do as well. Also Linux allows a user to transition to a big-iron UNIX machine much more easily, if they are used to using a UNIX on the lower end machines.
Slashdot Mirror
Actually, not anymore. Mandrake 7.0+ has branched off RedHat, and 7.1 is QUITE a bit different from Redhat. (It has ReiserFS, it has XFree 4.0, and the other packages are up to date.)
I did.
People who think that /. has every right to post messages before people have a time to mirror piss me off. There is a thing in this country called "responsible journalism." Since the journalistic community these days is in such a rut, I'm not surprised that most of you haven't heard about it. In traditional media, responsible journalism takes the form of corroberating your evidence to make absolutely sure that you're giving the correct news. Until recently, all respectable establishments did this. Those that didn't are called tabloids. Because of their rush to get the dirt, tabloids often post too early about a story. If it turns out to be true, then the tabloids will often have the news before any of the papers, but if it is false, then it ultimately hurts those that the story is about. In the new electronic media, journalistic responsibility has to extend to thinking about the ability of servers to take an increased hit before they are ready. It is analagous to allowing a defendant to prepare his case before the prosecution starts. In the good old days, the media had responsibility. They would rarely post a news story about, say, a president before he had made a public statement about the situation. The rumors and such were left to the tabloids. This situation is similar. You don't attack somebody (the /. effect is certainly an attack, though unintentional) before you give them a chance to prepare. It is just not decent. /. posting new files before the mirrors have a chance to gear up is not just impolite (or I'm I just a dinosaur and curtousy is passe?) it is also detrimental to the community, which wants to be able to download the software.
Reiser FS looks seriously cool. I tried it out a little on Suse 6.4 and it flies. However, there are a few things about it that piss the hell out of me. First, it isn't well integrated. I don't know why, but for some odd reason, LILO can't boot a reiser filesystem, so you have to make another partition for /boot. Its a pain, especially if you've got 3 OSs like me, and hit the 4 partition limit way way to often. I have to wonder, though, what the hold-up is. I'm not trying to troll here, but what's taking them so damn long? They've been working on it since .99 and its still not done? BeOS already has a from scratch fast, fully journaled FS that's been in there since at least R3. Also, it already has the database capabilities that won't make it into Resier for a while now. Why's the project taking so long? It's not like they're perfecting it, ResierFS still is less stable and loses data more often than BFS. Still, for Linux speed-mongers ( he he, oxymoron) Reiser is a must have.
Actually, a fully journaled filesystem need not be slow. Take a look at BFS. It is freaking fast, and fully journaled. (Hell, I've given up shutting down. Be should just remove the shutdown menu.)
Ahem, you don't even have to got that far, try BFS.
Oh yea, Linux is just so hideosly great. It can't match BeOS in response time (their midi kit now has microsecond response and only RTLinux can beat its audio latency.) It still doesn't have a GUI that is as elegant and inuitive as Be's, it is just now gettings direct access to graphics hardware (not just a simple frame buffer) and most relevantly to this article, it still doesn't have a journaling filesystem that is as fast as Be's, as stable as Be's, and has a built in database engine. Seriously, Linux is great and all, but before you spout bullshit, remember that there are a lot of areas (SPEED!) where it is not #1.
By opening up the playstation hardware, Sony has made a great "in" for itself. In the console market, nobody ever makes money on the hardware, they make money from software. On the PSX, Sony gets $7 for each CD sold. It only makes sense that if more people have conoles, Sony will make more money. As for the X-Box, I'm not to impressed with it. It looks like it will be a regular PC CPU paried with a regular PC graphics card, running a version of NT. Is it just me, or does this scream "innefficient." Face it, people are never going to want to edit documents or photos or use a console for normal PC tasks. The console will always be a reciever for content, wether it is from the web, a game disc, or from a DVD movie. Sure people might edit a few emails, but this whole integration of one stop photo editing, document writing, etc, is unlikey to happen. After all these years, your toaster is still seperate from your microwave is it not? Most people have both a toaster and a microwave, and I'm pretty sure that in the future, people will have both a set top box and a computer. By making the XBox so general purpose, MS does two things. One, it brings itself to the game late because of the extra development time. Only one company has ever been able to pull itself to the front after coming in late and that was Nintendo (then the 800lb gorilla of the market) with its SNES. Now MS is not the 800 lb gorilla of the console market. Sony is. By coming to the game late, they severely hurt themselves, because for a console, all that matters is quality and amount of software. Nintendo found this out the hard way with N64, and MS will likely get a beating from Sony because of this. Also, MS brings too much PC crap to the table. Face it, if NT isn't stable enough for your mission-critical server, it isn't stable enough for your console. The 300 days between crashes for a Linux box is megear compared to the 3 years my PS has gone without crashing, while being run at 100% by funky custom ASM code no less. The PC design also brings a lot of overhead. The PSX2 architecture is much more integrated than that of the XBox. Integration means a lot. Take a look at the N64. This thing has better graphics than a Pentium 200 with Vodoo and does full screen antialiasing and graphics features that PCs only dreamed about until the TNT era. It does all this with only a 97.5 MHz R4300i and a 66MHz graphics proc. If you take into account the efficiancy of the architecture, you'll notice that the XBox doesn't look so hot anymore. I'm willing to prognosticate that in the end, it will look like this. Sony comes out with the PSX2, makes a ton of money, gets a huge installed userbase, then a year and a half later, the XBox comes out. It is maybe about 30% faster, but doesn't have the huge gaming library, and is a year and a half behind Sony in building a userbase. In addition, people get frustrated by having to deal with PC-type technology (ie. crashing) and find another reason not to by the XBox.
UNIX is NOT user friendly if you don't do the things that UNIX doesn't like. Windows is not friendly if you don't do the things that Windows doesn't like. If rm -rf * is your style, the UNIX is your game. If right-click->Delete is your style, then BeOS is your bag. Windows is only your bag if you need to run a program. If you're running a server, good luck if you try to do it in windows. If your running a desktop, god help you if you try to run it on UNIX. I am quite experianced with Linux, but it still amazes me how much crap I have to put up with to do simple desktop things like install a new 3D card driver. (I'm not kidding by the way, it is ridiculously easy in Windows. More so in BeOS where the system automatically installs drivers.) Of course I've dabbled in networking, and am amazed at how easy UNIX makes it.
Really, you use an old mac to do website stuff? Increadible. I just buy new hardware. I am performance obsessed too, but I wasn' talking about a pixel level editing program. Neither was I talking about operating system-level stuff like GNOME. (Wether Linux people like it or not, for all intents and purposes, the DE is part of the OS.) I was talking about the 3D renderers and the Office apps, and the web browsers, and the photo editors and the video editors. The meaty stuff that people use every day. Sure there are those who have powerful hardware, but use older programs for the sheer speed, but unhappily, they are a rare case. My point was that developers know what kind of CPUs their stuff will run on (at least in the mainstream, I don't know about UNIX though) because people tend to run modern programs on new hardware. I am sure that the developer of that plain X app had no idea how powerful a PIII would be or how to optimize for it. But that's okay, because 99% of the time, nobody will use that old X app on a PIII. The point is that the compiler doesn't really have to worry about optimizing for different architectures from the future (in the same instruction set) because the program will rarely be used on those architectures. The post I was responding to thought that dynamic compilation would win out, because when the K8 or K9 comes out, software written for the older CPUs would be optimized better for these new CPUs. A static compiler can't do this, because it can only optimize for CPUs that are known at the time. I was saying that it doesn't matter, because people probably won't be using that app when the K8 or K9 comes out. Even then, the software won't need optimization, because it will have been writtent to run well on an older CPU.
I don't think you realize that you aren't talking about dynamic optimization, you're talking about optimizing compliers in general. A dynamically optimizing complier has a few edges over a statically optimizing complier, but in general, the static compiler wins. Think about the differences in hardware archicture. Aside from something like an OS, a developer already knows what CPUs their program will be used on. If I make a performance critical program now, say a 3D renderer, I know that it will be used on PIIIs, PIIs, Athlons, Willamette, and Maybe K8. That's it. People will not be still using my program on K9s and 10s. Sure some people use older programs, but think about it. In programs where performance really matters, do you use anything older than two years?
It would be a great deal faster since Visual C++ 6.0 Standard doesn't do any optimization.
The poor showing of Visual C++ was very surprising, given that Visual is held to be one of the fastest x86 compilers. The fastest x86 compiler at the moment is Intel's own x86 compiler and it would have been interesting to see how it would fare on the benchmark. I have done performance tests between Visual and Intel's on graphics apps on a PII, and Visual's is rarely more than 5-10% behind Intel's. A future extension to this could be to test this code with Intel's compiler since it is downloadable on the web (full but time limited) and easily pluggable into Visual. I suspect that the weakness of the Visual complier may have something to do with Microsoft not optimizing as much of a non intel CPU and the recent K7 optimizations gcc has gotten.
How 'bout the other way around? Think about it this way. There are no serious Linux 3D applications that can use the 3D accelerated drivers, (since most static link Mesa) so the only major 3D thing it can run at the moment are games! I'm sure this will change as 3D apps ar ported to Linux, but for now, Linux has nothing that wants to make me trade in MAX. Also, he was reffering to those running XFree 4.0 in general, not just the 3D stuff. Trust me, if you are a desktop user, you need XFree 4.0. I have not yet seen a crash and it is much faster than 3.3.x. In fact it restored my faith in Linux as an OS when I saw that finally XFree86 could blit stuff to the screen as fast as DirectDraw (windowed) could.
It's wonderful that SiS already has 3D support, but that wasn't exactly my point :)
This is awesome, except not really. It's great that a company is offering Linux support, but an announcement like this in the Windows world would be met with laughs. SIS has never made even a mediocre graphics chip. All their chips have rated from Man this sucks to Man this sucks as much as Windows 2000! It really shows how easily exited the alternate OS community is about any support whatsoever.
KDevelop alone is worth the price of admission (which of course is $0.)
:)
I tried it last winter, and found it utterly useless for creating anything but KDE apps. I'm
trying to write ISO C++ libraries, and Kdevelop (at the time, at least) didn't even have
a library project template! Maybe I'll try it again sometime, but I was seriously
disapointed with what I saw.
>>>>>>
Huh. If I'd tried GNOME at that version level, I'd have run away screaming. Try anything 1.1 and above. It is rock solid, and now (in 1.2) has integration with GNOME development files. It is a serious competitor to VisualC++ if they can only get rid of that annoying "I" beam, and get IntelliSense type stuff in there. I develop entirely with classes and it is a bitch remembering function headers.
This is not meant as a slam towards the Kdevelop people, for creating a KDE app it
would be great. It just didn't do what I needed it to do.
>>>>
Try using an app when it is release-quality. I don't judge Linux on my impressions of Slackware 3.5, and you shouldn't either.
BTW, why did you mention Kdevelop, when you claim you are talking about the general
desktop market, or as you put elsewhere: "Linux grass-roots zealots weren't invited.
>>>
I was talking about the general desktop market. Many programmers are part of this too. Real grass-roots zealots use VIM
Correct. However you conveniently ignore the fact that the reverse can also occur. Will it? I don't know. But it's about as likely as your scenario is.
>>>>>>
Not really. There is a thing called a landslide effect. Once the tide significantly begins to turn in one direction, it is unlikely that anything short of a revolution will overturn it. KDE is significantly ahead in the apps and developers game, and it doesn't look like GNOME has anything on the horizon that will turn the tide. (Evolution will only take you so far.)
Have you by any chance used Gnome 1.2? Didn't think so. Go use it. Then come back and say that again. Or don't, if you suddenly find that statement you made holds no water whatsoever.
>>>>>>>>
I have, and my statement still holds water. GNOME 1.2 is still slightly slower than KDE. True, it is better in this release, but here is a good test. Take an app that has a lot of views (like ksysv or whatever the equivilant GNOME app is) and resize it. Watch the redraw. See which one redraws faster. GNOME 1.2 is about on par with KDE 1.x in terms of speed, but KDE 2.0 is actually faster than KDE 1.x on my machine.
You're some sort of psychic, I take it? It seems to me as though Gnome and KDE are making the exact same types of changes along their paths to 2.0. New file managers, component architectures, themeability on KDE's side, etc. Gnome's added a 1.2 step, but what's the harm in that?
>>>>>>>>>>>
Yes, I'm phsycic. My point is that KDE 2.0 has less memory usage than GNOME 1.x. Since KDE 2.0 already has all the new stuff in place, it is doubtful that it will grow too much in the future. GNOME 1.2, however still doesn't have all the features in place, and can do nothing but grow in 2.0. Thus, KDE 2.0 will take significantly less memory than GNOME 2.0.
And you speak from... how much experience? As I thought, none. Go use it, then come back when you can make arguments and back them up. If you can't back up your arguments, you're just flaming.
>>>>>>>>>>>>>>>
I don't speak from firsthand experiance with either environment, true, but I can tell you this. Bonobo is similar to KDE's original object architecture, which was based on MICO. They ditched it in favor of the much simpler and faster DCOP arch. I doubt there is much the GNOME guys can do to make an inherently slow architecture (even the Berlin guys admit this) fast.
Ah, and thus we get to the whole crux of your argument, and the real reasn you're posting: "KDE is better because I'm a Corelite." Let's see your examples. I hve no doubt they probably exist, but where are they? You're certainly not too cooperative in pointing them out. Why not?
>>>>>>>>
My first Distro was Slack. My second distro was RedHat 5.1. My current distro is Suse 6.4. I'm running KDE 2.0 beta 1, on kernel 2.3.99pre8 with the beta nVidia OpenGL drivers hooked up to XFree86 4.0 all running on ReiserFS. I am certainly not a Corelite. Even an advanced user appreciates cohesion in their work environment. Plus, it has nothing to do with advanced users anyway. My discaimer at the top got cut of for some reason, but I wasn't talking about the "core" linux crowd. I was talking about the new Linux users (ie. the business) market, those who couldn't care less about GPL vs. QPL and only wanted a fast, stable, COHESIVE desktop. If you use Linux because you love it, then this post doesn't concern you.
According to whom? KDE has StarOffice, but other than that every KDE app I've seen has a Gnome analog. For the most part, the revere is also true.
>>>>>>>>>>>>
KDE has nothing to do with star office. KDE has KOffice, which by itself gives it an enormous lead over GNOME and whatever it's office suite is. (A mish mash of Abiword, Gnumeric, Evolution and a few others.) Aside Evolution, KOffice looks like it will blow the others away. Then, of course, there is KDevelop which again is a "killer app." Both platforms may have similar apps, but in general the KDE ones are higher quality/more useful.
Again, you conveniently ignore the fact that the people could just as well choose Gnome as KDE.
>>>>>>
My whole post is about why people will choose KDE over GNOME! This point only accentuates the fact that people (most) will choose ONE, not live with both.
Which is...? Your answer is conspicuously absent. As if you realized that since you have near-zero experience with Gnome, particularly recent versions, you just might be wrong. Use them, if only for a little while, so you can make a real comparison in that regard.
>>>>>>>>>>
KDE has more and better apps. I've been over this already. KDevelop, KOffice, and the great set of apps bundled with KDE are much better than their GNOME counterparts. Do I even have to debate the merits of KDeveloper of gIDE? I have a lot of experiance with GNOME in fact. I've been using Helix GNOME on RedHat ever since it came it. I had used KDE for a month before that, but liked GNOME's asthetic sense much better. Near the end of my RedHat cycle, I tried KDE 2.0, and switched to it and Suse 6.4 a week ago. (And yes, I do have both Helix GNOME and KDE installed.)
And please don't forget that even if you are correct, that means little. Remember, people stuck with Windows, which hardly has the "best" apps (the single possible exception being Excel, and both Linux DE's have spreadsheets which are coming along quite nicely in that regard).
>>>>>>>>>>>
Are you kidding? People stuck to Windows because it has the best apps. Where it doesn't (ie the high power server market) people don't use Windows! Are you telling me that Gimp beats photoshop, or anything on Linux can beat MS Word, 3D Studio MAX, Maya, SoundForge, Cakewalk, Truespace, IE (it's true, Netscape sucks), Visual Studio (if you're into the IDE thing), and the limitless games available for the platform? These are best of breed apps and are ONLY available for Windows. People don't use Windows because MS forces them to. People use it because it has great apps. People like me reboot into it because they have to use 3D studio or photoshop. (Because they rock!)
For better or for worse, Gnome is just as Windowsy as KDE. And both are getting even more Windowsy as the versions progress. It's quite sad, really. You'd think the Open-Source community could do better, interface-wise. It's been done. NeXT had a better GUI which was almost completely original. So did MacOS. Win9x and OSX are fusions of these two; Win9x took the worst of both and added some decidedly anticompetitive elements. It'll be interesting to see if OSX has it done right (I admit I don't like some of what I see, but I'll reserve judgement until I've actually used the thing).
>>>>>>>>
You miss my point. I wa critisizing people who think KDE is too Windowsy. I neglected to mention GNOME, but sure, why not. People could really care less about the quality of the interface. People like something that is familier, no matter how hard it is to use. (Of course, they won't bother to familiarize themselves with something that is harder than what they are using now, but also won't switch unless the competitior is revolutionarily easy. People are lazy.)
Show me your "growing support from the business community." I don't see it. I saw StarOffice getting KDE integration back when it was still proprietary (versus the pseudo-proprietary Community-Source liense it's under now). Oh, and Corel bundles it with their system. That's all I ever saw. If there's more I'd be glad to hear about it, but I simply don't see it anywhere.
>>>>>>>>>>>>>>>
Oh, and Corel bundles it? Like that is a little thing? Corel is the only respectable (traditionally) company doing Linux, and you act like it is a small deal? Also, must Linux distro's have much better support for KDE than Linux. Go to Mandrakes website. Compare the number of GNOME screenshots to KDE screen shots.
Give an example of a "big, memory-hogging" feature Gnome has that doesn't make it better.
>>>>>>>
Bonobo. It makes it better, but not enough to warrent the weight.
Depends. Everyone knows that Gnome 1.0 was released too early; I can't dispute that claim. It was underfeatured and unstable. The stability got fixed quite some time ago. 1.2 adds more of the features. Not all of them yet; a few are still in development (Bonobo and Nautilus, most notably).
>>>>>>>>>>>>
It took GNOME until 1.2 to become as stable and nearly as fast as KDE was almost a year ago. What does that tell you?
Speed... well, I'm not so sure. A lot of that seems to have to do with Imlib. Everything that I've seen that replaced it wuth gdk-pixbuf seemed to get a big kick in the pants, speed-wise.
>>>>>>>>>>>>>>
Great. 2.0 will be faster. But so will KDE 2.0. At everystep, KDE seems one step faster than GNOME.
Of course, you can't expect phenomenal speed when you're drawing your entire GUI with pixmaps. That's a large reason of why Gnome feels so slow for a lot of people. Switch out Sawfish for Window Maker and the Pixmap themes for GTKStep and you've suddenly got a much faster desktop. I like Sawfish myself, and used it for several months, but as of this moment it just doesn't seem to be able to touch Window Maker for speed and stability, though Window Maker's icons do get in the way on a Gnome desktop. Things may have changed with the developmental gdk-pixbuf versions of Sawfish; I intend to try it out as soon as I manage to compile the thing.
>>>>>>>>>>>>>
KDE 2.0 can draw the entire desktop with pixmaps much faster than I've ever seen GNOME do it. Plus, my themes to use pixmaps, so very little is being drawn with pixmaps.
Either way, this is what's wrong with the state of discussion on Slashdot. People don't debate anymore; they just flame. They'll spout drivel like this post's parent, which say a lot but prove nothing, and thus are no better than "Gnome sucks! No, KDE sux0rz! No way, Gnome 0wnZ KDE! No, KDE is ph4r m0r3 k-r4d @nd l33t! No, you must PH3AR GNOME!" What happened to people not jut making their point, but backing it up? OK, so Gnome sucks NINJA ass; why? OK, so maybe KDE sucks big fat naked petrified donkey dick; why? The point is, unless we get into real discussion, and point out the relative strengths and weaknesses of the two systems, neither one is going to truly improve. They'll just get stuck in an endless feature war, neither one getting the features users really want. They'll just try to one-up the other, and the end result won't be any better than Win2K.
>>>>>>>>>>>>>>.
I just pointed out a dozen weaknesses of GNOME. I'm not saying KDE is flawless, but GNOME is still much worse. I'm sorry if that's not what you wanted to hear, but for the mainstream market, (my disclaimer got cut dammit), KDE is still better and getting better faster than GNOME is!
My account info got lost so I m posting as an ac (e-mail mike@redtux.uklinux.net)
>>>>
Sure
Points in no particular order 1.
KDE is faster? In my experience KDE has never been faste. I recently installed mandrake 7 on my work box with
kde 1.1.2 and october gnome. KDE is still far, far , far slower than gnome to to do anything useful. I don,t care
how fast it redraws the screen, ftp transfers in terms of bytes in a term I do care about.
>>>>>>>>
Doing the French add an 'e' to every word thing? Are you a desktop user or a server admin? Of course screen redraws matter. Whats the point of using Linux if your GUI is going to be slower than Windows? And what does the DE have to do with FTP transfer? If anything, it would be faster under KDE due to fact that the ftp program would have more memory left.
2. Gnome unstable
yeah!! nuff said
>>>>>
Face it. GNOME stability sucked until 1.0.50 came out, and even then it was crappy. The first stable GNOME (by Linux standards) was October GNOME running a WM other than enlightenment.
3. Better integration of KDE apps. I use Linux - I do nort want a monolithic desktop.
>>>>>
Did you read my disclaimer? Linux grass-roots zealots weren't invited. I was talking about KDE and GNOME in terms of a viable desktop environment, not something that fuels some nerd's desire to be 'leet just because his apps don't match.
BTW I
stopped using KDE at 1.1.1 because it was so slow and buggy and went back to fvwm and then started using
gnome at 1.0
>>>>>>>
The sheer fact that you'd fvwm shows that you aren't a desktop user, and the fact that you think GNOME 1.0 was anything but alpha software (again by Linux standards) shows that you're on crack.
One of my points was that it was all about the apps, and KDE has the good ones. Though the Gimp is probably irreplacable, KDE has the majority of the good apps. KDevelop alone is worth the price of admission (which of course is $0.) In addition KOffice is shaping up startingly quickly, and though the spreadsheet may never be as good as gnumeric (doubtfull, as the KDE guys seem to know their apps pretty well) the rest of KOffice will be a major factor. Also, KDE has a lot more developer support than GNOME does at the moment, and in general the KDE apps are far more mature than their GNOME counterparts. (Aside from a few exeptions like Gimp and Gnumeric.)
Every time I see another release of GNOME, I shudder to think of what could have been. At this point, it is obvious to me that GNOME will lose the war with KDE in the desktop environment department. Now don't start on me about how free software cannot lose a war, or about how the two are not competing. The one that loses is the one that has the fewest users and a declining userbase. Since both DEs have their own API, a DE can lose the battle if nobody develops for that API. If most people start using KDE, then few people will bother to develop for GTK/GNOME, thus GNOME will have lost the battle. You can say that GNOME is a project and will continue wether or not it has support, and you'd be right. A project can still be alive and still have some users and still have lost. At some point, it just become irrelevant. Take a look at OpenStep. Sure it may oneday rise up and grap the hearts and minds of Linux users, but it is more likely that it will continue to putter along, maybe contributing some valuable technology to one of the dominant DEs in the proccess. Even if the developers do not want it to be that way, GNOME and KDE are at war. They are at war for the deskspace of the user and the coding time of the developer. Up until recently, the only reason I got anywhere near GNOME is because it had the nicest user environment available in Linux. It was asthetically pleasing, and generally nicer to work with than KDE. Of course, in KDE 2.0, KDE took GNOME's asthetics and usability and made it its own. Now that both are on the same level, interface-wise, KDEs strenghts can come out and allow it to pull ahead of GNOME. In a succinct list, these strenghts are,
1) Speed and resource usage. KDE 2.0 is by far faster than any version of GNOME, and takes a significant amount less memory. This disparity will only widen in the future, since the general GNOME infrastructure will grow quite a bit in 2.0, while KDE's probably won't grow much beyond 2.0 since the basic functionality is in place. Couple this with the fact that DCOP/KParts is far superior to Bonobo in terms of speed and resource usage (though maybe not as flexible in a distrubuted environment) KDE bowls GNOME over in the speed/resource usage deparment. This is significant becuase people don't leave Windows to arrive at another bloated OS.
2) KDE still has a much tighter integration between apps than GNOME does. It only takes a cursory look at Corel Linux to show what a little ingenuity and KDE can do to make Linux almost as friendly as Windows for the desktop user.
3) KDE has far better apps and much more developer support. Sure the KDE libraries can be loaded, but with the increased bloat in both 2.0 level libraries, people will increasingly wish to not have to load the libraries twice. As such, they will stick to the DE that has the most/best apps.
4) KDE is more familer. Diehard Linuxites call it being too Windowy, but it is a big strength for the rest of us who grew up on the start menu and Explorer. It is a moot point which one is theoretically easier/more intuitive. If that mattered, NeXT would still be around (or AfterStep would be dominating in the DE department.) Whatever is easier is whatever people are used to. MS might have the clout to force a new style of desktop on people (a la Windows 95) but Linux does not.
These factors, coupled with increasing support for KDE from the business community will allow KDE to become the dominant Linux desktop.
Back to my thoughts about GNOME, it saddens me to think of what GNOME could have been. It has a great set of people working on it, and the developers have a great asthetic sense. However they seem to think that putting the biggest, most memory hogging features into the DE will automatically make it better. Software size and quality seems to not be a very top priority, and software speed seems to be an even lower priority. Features seems to head the list, and everyone knows what that mentality gets you... it gets you Windows 2000.
It could, but PCMCIA support is quite a bit easier. Also, the kernel still has to have some sembelance of what PCMCIA is (in this case a bus) in order to load the module. Sure DirectX could be loaded as a module, but it would still need a great deal of access to hardware without going through the kernel, and would still have to be an integral part of the kernel. It might be loaded from a module, but the kernel would still have to include new code to allow the module to take over the system. (What do you think DX is for? Total control baby ;)
DirectX would have a hard time being ported into Linux. Like NT, Linux has a significant admistration layer, and it wouldn't meld to well with the rest of the OS. Unlike OpenGL, which has a higher level interface to the OS, DirectX has to be down there in the core. Unless the Linux gods officially bless a DirectX port and integrate it deep into the kernel, it will be hard to get a good, fast DirectX port going without forking the kenrel.
Now for my obligatory BeOS spiel. Wine is not only helping LInux and other Unicies, but alternative OSs in general. In particular, Wine has already been compiled on BeOS, and things are moving at a fair clip to get the thing working. Wine is really a godsend to some of these alternative alternative OSs because it will allow them (when it's done) to have a body of functional software that allows users to complete some important tasks. Unlike Linux, these OSs are missing a few critical apps, and what better way is there to quickly fill this gap than get good emulation (I know it's not emulation per se, but for lack of a better word) going so Windows apps can be used. If you have any coding talent, support the BeWine project at bewine.loungenet.org. It seems that it is pretty low profile and not enough people know about it.
I wasn't talking about PC OEMs. I was talking about the lower end workstation OEMs like Compaq, SGI, IBM, and HP that have high end UNIX machines, but had adopted NT for its lower end, especially workstation lines. Moving to Linux for these machines gives these companies much more flexibility and power than if they had used NT. I know that sgi hated using NT for its machines, and I'm sure Compaq, IBM, and HP do as well. Also Linux allows a user to transition to a big-iron UNIX machine much more easily, if they are used to using a UNIX on the lower end machines.