What if FreeDesktop.org started to become the “free software desktop project” and GNOME and KDE slowly became “flavours aimed at different audiences”, Seth Nickell wonders in his blog. Seth is having some good ideas about how the future of X11 DEs should feel like, while down that page there is an extra explanation of Storage, the technology we reported last week that Seth is developing.
How KDE & GNOME Fought their War and Got Things All Backwards
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I just can’t take seriously the opinion of KDE enthusiasts. Trolltech can talk to me after they’ve published some sort of QT HIG that doesn’t suck. Until then, I’ll be using GNOME, which appears to have been designed by people with actual knowledge about interface development.
Also, his analogy regarding the universality of the “desktop interface” is laughable. No, true, there is no “one correct” car for everyone. However, all cars possess the same interface (at least to an extent)! The whole point of an interface is to allow people to apply previous experience to the current situation.
KOMPRESSOR
>I just can’t take seriously the opinion of KDE enthusiasts.
I don’t know why you think that, but Seth is a Gnomer…
It appears that I was misled by his comment about being able to recommend KDE to someone. So scratch that first paragraph, although I still am contemptuous of KDE.
However, I think my second point is a valid one.
KOMPRESSOR
besides knowing some real car enthusiasts that would beat the daylights out of you for saying they all have the same interface…
why are you railing against kde? Gnome seems to be all over that blog of Seth’s.
am i missing something?
so the only thing we have to discuss so far is why kompressor really hates kde.
i’m just curious. i use gnome predominantly, but i find that not mentioning the window manager is really a shame. it’s half the experience.
and i’m writing this from KDE right now, because i’ve had it up to HERE with metacity.
so i put sawfish back on, and blech. sawfish & gnome…i’ll stick with metacity and gnome.
but for now, i’m using KDE. i finally got my wire-frame mode back, i’m tired of opaque windows being forced down my throat.
if it’s not double buffered, and smooth as butter, take that shit out.
gnome/metacity opaque windows being dragged around and or resized look like a complete joke even on the fastest systems.
I just can’t take seriously the opinion of KDE enthusiasts. Trolltech can talk to me after they’ve published some sort of QT HIG that doesn’t suck. Until then, I’ll be using GNOME, which appears to have been designed by people with actual knowledge about interface development.
Also, his analogy regarding the universality of the “desktop interface” is laughable. No, true, there is no “one correct” car for everyone. However, all cars possess the same interface (at least to an extent)! The whole point of an interface is to allow people to apply previous experience to the current situation.
Just a quick pointer, the GNOME HIG was a joint venture between SUN and the GNOME foundation. SUN paid and conducted the “tests” which provided the necessary data for the development of the current HIG.
Unfortunately, however, there is no standardise HIG. IMHO, KDE should just swallow their pride and and adopt the GNOME HIG, we’d be all better for it.
KDE is largely veering towards the “Power User” market because for so long they ignored usability issues like the massive amount of menu clutter and options that violated Hicks’ Law in every known way possible, making things so difficult for anyone just starting off that only a geek for whom a computer is the ends and not the means (aka “Power User”) could actually tolerate the environment.
GNOME is largely veering towards the corporate desktop because a programmer ripped out every single option imaginable without thinking about whether the option or feature might actually make sense if it was intelligently laid out and the necessary progressive disclosure was appropriately applied.
A war between GNOME vs. KDE is like a war between Sauron and Saruman; it doesn’t matter who wins in the end, because both are equally capable of making Middle Earth really suck.
I’m not impressed. Regardless of how they are buffered (and buffering is a function of the windowing system AFAIK, not the widget set) the contents of KDE windows look like ass. Inconsistent layout, inadequate ease-of-use consideration, etc.
In any case, I only made the first comment because I misread and thought that the article was advocating KDE. I personally feel that the “success” of KDE has put Linux far behind in the competition for the corporate desktop. However, the point only came up because of a misunderstanding on my part.
I’d much rather talk about the second point, which is the fact that the author’s analogy is patently ludicrous.
I suppose I must amend my statement to say that NEARLY ALL cars have approximately the same interface. Which is to say, the interfaces of the cars provide nearly the same affordances: you may depress the accelerator or the brake, you may turn the wheel to make the car turn, you may operate the shifter (of which I must confess there are two types–standardized, but different nonetheless). Also note that cars that don’t possess an automatic transmission possess a clutch, but the clutch always performs the same action. There are no cars controlled by joysticks, and the gas and brake pedals always appear in the same configuration relative to one another.
This is what I mean by having approximately the same interface–not the degree to which the vehicle is responsive, etc. The uniformity of the “car interface” permits me to hop in just about any vehicle I please and begin driving. Notice that, even though standard and automatic transmission cars are forced to have different interfaces by implementation details, safety-essential elements such as gas/brake/steer remain the same. This fact means that the interface, once learned, can become almost instinctual–because you absolutely must not pause to think sometimes before swerving “now how do I go about making the car turn?”
People seem to fail to understand the importance of a uniform interface. Linux will not penetrate the corporate market (and we won’t get the goodies that come with that penetration) until there is a consistent interface. Why? Because any time spent trying to figure out how to make the interface do something is time NOT spent making the interface do that thing. Companies aren’t interested in downtime–the $100 for a seat license of XP or whatever is peanuts compared to the (added up) time users spend trying to figure out unfriendly interfaces.
This is getting kinda ranty and I’m getting kinda tired. I’m calling it a night.
KOMPRESSOR
i read OS News probably every couple days just to see what’s going on and it’s amazing how fast the quality of comments has gone down in the last 6-8 months. =( i suppose that means that OS News is becoming successful and has drawn a more general audience.
first, Kompressor: Seth is a GNOME developer; take your stupid project hate somewhere someone might actually care about it (i suggest your momma). i don’t agree with all of Seth’s viewpoints, and i think his storage project has some “challenges”, but i agree with him that freedesktop.org is an important project that is becoming more and more a vital part of the Free Software desktop landscape. and i happen to be a KDE developer. OMG! people from the two projects agreeing on something? i hate to break it to you: but it happens all the time in this modern age of miracles small and great.
i do think seth tended to trivialize and oversimplify the point at times in his blog (look at the struggles to get an audience for DBUS for something as trivial as local IPC, for instance), but what’s a blog for if you can’t be starry-eyed and engage in stream-of-consciousness rambling? some of the goals he states are achievable sooner rather than later (IPC, XML parsing, media framework), others not so (HTML rendering, mail transports) due to technical preferences and decisions. but the core of the concept is sound and is playing out even now.
i find it amazingly humorous how many of the comments here so far have missed his point.. anyone who rags on either GNOME or KDE in this thread needs to re-read what Seth said, as many times as necessary to get it through your head that it’s about finding pathways of cooperation and creating unique visions upon that cooperation. perhaps this is exactly what is needed for Linux to achieve the fabled (and tongue-in-cheek) world domination of the desktop.
as for “KDE swallowing their pride” and accepting the GNOME HIG, well… perhaps if you actually understood what the GNOME HIG contained and looked at the two desktops you’d understand that it has NOTHING to do with pride. thanks for the insult, though. >:-( btw, watch for a new and improved KDE UI Guidelines document in the coming year.
Besides all this corporate talk we shouldn’t forget that GNOME is still communitywork with many participants. It’s mostly SUN, Ximian or RedHat (and their employees) who like to see GNOME maturing to a corporate Desktop. But many participants have no concrete idea what this all embedds. People follow it without knowing what the end result looks like. While Seth is an important person in the GNOME community with all kind of ‘ideas’ we shouldn’t forget the many other people involved into it who have ‘ideas’ as well which not necessarily match the ideas of Seth or Havoc. Ignoring them can be a dangerous thing because they are the ones developing and contributing to it. KDE and GNOME have both good points as well as bad points. There is no need to take the one to the shambles and hyping the hell out of the other. Bringing up stuff like ‘GNOME is the corporate Desktop’ is based on what facts ? (I mean, we see it moving towards this direction) but the last word is still been spoken by the community and participants. I would pretty much like that people understand this.
IMHO the biggest problem of KDE future the QT. You cannot develop KDE based commercial application, because QT open editation and commercial editation is not binary compatible. You can create Qt application (but QT is very expensive solution), but not KDE application. Gnome is far better for developers. If Gnome can integrate the Mono or DonGNU project and anybody create a good RAD/IDE tool for C# it can be comparable environment to XP and .NET.
A war between GNOME vs. KDE is like a war between Sauron and Saruman; it doesn’t matter who wins in the end, because both are equally capable of making Middle Earth really suck.
I couldn’t explain it better!
😉
H.a.n.d.
Golum at work dot ork
I think your statement has a weak sidenote. If you develop commercial applications with QT then you can easily sell your app for some calculated money to return your investiment (ROI) of what you paid QT’s license for. I mean when you develop commercial apps then you plan to sell them and make cash for it. On the one side you want to make cash for your app but refuse to pay cash for the toolkit you use.
The MONO issue needs also been viewed from a different point. MONO is being developed by Ximian and making it become part inside the core of GNOME can be a dangerous thing to do. Not every participant would agree this to be a good decision. If you have conversation with a bunch of people in the GNOME community (which you can easily do by participate on the gnome IRC channel or throughh private chats) then you figure out that the people have different interests and different opinions.
MONO may be an interesting projects but I recommend it to stay Ximian’s baby for the next couple of years until the market has been researched properly and the community have been asked what they think about it before making the stupid decision to put it inside the core of GNOME and whipe away a bunch of important developers who may NOT agree with this idea.
This community is a delicate one. If there were corporate interests behind it e.g. one company driving GNOME then forcing MONO inside it may be possible but GNOME is and hopefully stays communitywork. We don’t want to disrespect people’s work here or decide over peoples head (those who contribute). I for my personal own don’t like MONO. I don’t say it may have no fugure. But it’s a personal choice. My choice to dislike it. Putting it in GNOME as core may me scare away, thus leads into not contributing to GNOME anymore, stopping my own GNOME related projects. This would be really sad for the community.
lets try again tomorrow.
for the record, i like both gnome and kde.
QT: It is true, but the price be commensurate with the features. I have commercial developer tools (Kylix Professional, Delphi, JBuilder) but IMHO the QT is too expensive. And I cannot develop KDE applications with this toolkit, only QT applications.
Mono: I not insist to mono, DotGNU also good thing, but IMHO the C# and managed code will the future in desktop applications.
LC: i’m not sure what makes you think that the commercial and Free versions of Qt are binary incompatible, because they are the exact same codebase. you can indeed create closed source KDE apps, and people have actually done this in the past. so i’d suggest that you stick to commenting on things you actually know something about, rather than spread FUD about vital and important projects.
some of the people on this board astound me. =(
Okay. I’m a bit tired of people complaining about the QT licensing. To license QT Professional for one developer is $1500. The cost per user descends after that. Any company creating software products is likely to spend well above this per programmer per year in keeping their hardware up to date. Assuming your coder costs $35,000 a year minimum for someone who knows what they are doing, the QT license cost is a drop in the ocean. The productivity gains from using a quality toolkit outweigh the price by far.
The thing is this; QT allows you to both produce free software and commercial software. BUT you aren’t able to produce low-commercial software, of the type that might only gather a few hundred dollars back. To be frank, it pushes you towards making the software a community project and free, and that is a good thing.
Most shareware won’t easily work with QT. This is a good thing. Do a Google search for ‘windows webcam free software’ sometime and dig around a bit. It is hard to find a piece of webcam software that will upload images to an FTP server that is actually free. It’s not hard under windows. You just grab the frame from the Video4Windows stream, present it, open an FTP connection using a pre-built web component from someone else and send the image. It’s a few hours of work, and yet when you look for a free version, you get shareware and free trials of applications charging up to $150! Heck, even windows screensavers most often come with a price tag. I am very glad that the QT license is preventing this from happening with Linux.
To sum up, QT will let you:
– Create a great free application for the community
– Create a great application to use within your business, as long as you don’t sell it.
– Create good retail software
It won’t let you:
– Produce small pieces of software and charge the earth for them.
So, to be frank, QT’s licenses work well and help to nudge people towards playing well with the free software community. It’s a great thing!
Oh, and for the reference of all of those people having a go at KDE’s context menus and things and saying that the developers don’t care about usability, Aaron is working very hard on it and has had a major hand in 3.2, which is bringing a massive number of usability benefits.
And final food for thought, a lot of Gnome application developers avidly dislike the Gnome HIG. They don’t like that Havoc’s vision has overtaken the project and they dislike adopting the HIG, and in places, find that they can’t implement it, because it breaks things badly. <- Not my opinion, but from a friend of mine who is an active Gnome developer.
Thank you for a level-headed comment. Although I still don’t see this “GNOME forcing”, as if there were some group of faceless individuals, convening in dimly-lit rooms, plotting and scheming, in sinister ways, to thwart potential GNOME developers.
MONO will never be “forced” down anyones throats- no one will ever “have to” use MONO to contribute to GNOME. MONO is but one of a dozen language bindings for GNOME. It offers certain features which are not present in other software and it is much easier for coding complex applications in which large amounts of information are shared between different applications. I too have my doubts about what microsoft may want to do with .NET. I am certainly not a fan of .NET. But if MONO can implement its own winforms it will gain its independence from microsoft’s implementation details. Of course the goal is to be able to create cross-platfrom software easily, yet GNOME is NOT cross-platform, neither in its implemenation nor in its goals, so the cross-platform issue is more pie in the sky. Nevertheless if MONO can implement enough of .NET that certain applications written for Windows can be painlessly ported to linux- well the more the better. MONO has made amaing progress, although its code base is still in its infancy it is already being successfully used, and Dashboard is showcasing this technology now(although it still needs 6 months prior to being viable).
Seth’s comments were very, very reasonable. I, for one, am really looking forward to trying out storage when it comes to fruition. Freedesktop has a huge task ahead of it- If only the bogus KDE-GNOME wars could finally stop, real progress could e made in cooperation and shared-code where such is practical. D-BUS is an excellent idea, several KDE folks have embraced, while others are still fighting an ideological cold-war which should already be history. I cannot see any point in the future where I will use exclusively GNOME or KDE applications-I prefer GNOME, precisely because of its sparse aesthetics and simplicity, yet I understand and value the wealth of configurability and modular structure of KDE. Users will decide, and decide daily, which Desktop they prefer, primarily dictated by their needs-certain tasks require KDE’s modular approach, other tasks don’t.
Linux development can not be coordinated and controled. If that happens in the future, I am going to quit Linux and go
looking for something else.
DG
Just thought I’d mention about DBUS… This is what I have taken from the mailing lists, so usual rules about taking it with a bowl of salt apply.
It’s not that KDE developers aren’t willing to take it up, just that DCOP and DBUS do slightly different things, and where they meet in functionality terms, DCOP tends to be a bit better. It is fair to say that KDE developers don’t want to adopt DBUS and scrap DCOP, taking a step back in functionality, at the same time, they see it as being able to augment existing DCOP systems.
DBUS operates at a lower level than DCOP in many ways, but the two can work together more or less in harmony.
Efforts are planned to produce a DCOP->DBUS bridge, so that should resolve the issue quite well and allow applications to interoperate.
Hope that helps to clarify things
We, the users, don’t care a rats ass about GNOME HIG, MONO or DCOP.
Repeat: we couldn’t care less. We don’t have the slightest interest about them. Most of us don’t even know what they mean.
This we care about: We install linux. Easy. We find lots of free apps, the fact that anyone will like. Everything is there, for free.
And we also care about this: some apps look different than others. That sucks, it really really sucks! A user at the KDE control center sees font setting dialog and thinks: “Whee! I can make the fonts smaller! Now the menus won’t fill up my entire screen!”
See the user opening mozilla, or gtk app, or whatever of the myriads of widget set hordes there are. Are the users font smaller, as he recently set them in control center? Does the user care about GNOME HIG, MONO or DCOP when they are not?
“The thing is this; QT allows you to both produce free software and commercial software. BUT you aren’t able to produce low-commercial software, of the type that might only gather a few hundred dollars back. To be frank, it pushes you towards making the software a community project and free, and that is a good thing.”
I think the issue for distro vendors, especially redhat (targeting corporate desktops), is that they cannot put together a system and say that if you want to create non-GPLd desktop applications for it, you need to go to a third party vendor and buy developer licenses at $1500 a pop. Thus they standardize on Gnome instead.
The point where your thesis breaks down is that most distros have standardised on KDE. I guess they didn’t agree with you that they had a big problem.
Rich.
If basic things like keyboard shortcuts, cut/copy/paste and drag and drop worked consistently between apps I could easily live with cosmetic differences. Surely GNOME and KDE developers can standardize those basic features without either of their projects losing anything?
These are consistent in Windows, Mac OS and just about every other GUI I’ve used in the past 15 years. This lack of consistency makes Linux a less productive desktop OS and makes it feel amateurish compared with Windows. Fixing this should be a top priority for both KDE and GNOME, it’s a lot more important than adding graphical glitz IMO.
I just can’t take seriously the opinion of KDE enthusiasts. Trolltech can talk to me after they’ve published some sort of QT HIG that doesn’t suck.
I would do the same for GNOME once GTK release a GTK HIG that doesn’t suck. This is KDE’s job, not QT. And they already have a unupdated HIG
I thought, for a long time, that HID was a simple science. You made things easy to do, and not confusing to the end user.
Then I started taking CMSC434 – Human Computer Interaction (HCI) – and started doing some reading.
It is now my opinion that most people who criticize the GNOME HIG are usually wrong, and more to the point, aren’t qualified to criticize in any meaningful fashion in the first place. The amount of literature available on HCI is staggering, and it’s more than “make the buttons big”.
Why don’t some developers like them? Because they actually make them think about users other than themselves.
The first rule of HID: there’s no such thing as user error in operating something. If you don’t understand that, I don’t think you’re qualified to whine about the GNOME HIG.
-Erwos
Perhaps I am somewhat confused. I used KDE on one of my systems. On that system, I have QT Commercial version. We make applications using QT for inhouse (we make money with them, so they are commercial). KDE works fine with the commercial QT technically, is there a license issue with what we are doing?
SMF
Seth makes a point. I have always been disappointed that freedesktop.org didn’t take of faster….
Surely there is loads of standardisation that could be done across both desktops which would not ‘interfere’ with each other? Mind, it’s not only on a desktop level but also on an application level: If I happen to be in KDE and get my pop mail with KMail, why can’t I see it when later I open up Evolution (or Balsa or whatever…) Why not have a standard location/format for mail? The Office packages SEEM to be going that way (with KDE, Abiword and OO.org all migrating to using the same format) Browsers. We have lots of pretty good browsers in Linux. Why do I have to set the same setting over and over for Mozilla, Konq, Opera, Galeon, whatever? Shouldn’t there be a common baseline of functionality (cache location and size, proxy info, Nescape plugins etc) Same for news readers.
The great variety of application to choose from is one of the biggest draws to Linux for me. Abstracting the ‘variable end-user data’ from the applications would seem to me to be a step forward.
Funny. In the beginning, I had the feeling that KDE was the button down, ‘corporate’ desktop while Gnome was the Linux enthousiast’s dream with loads of things to tweak and change (Enlightenment didn’t help that impression….). Now Gnome, with it’s good, highly adhered to HIG seems a lot more buttoned down, while KDE as of version 2-3 has become eyecandy goodness in addition to being a great all round desktop.
The mail inconsistency is not a unique GNU/Linux Gnome/KDE thing. Try opening your pop3-mail in Mozilla/W32 and Outlook (Express). Same thing – different formats. If you want to share mail across applications, you must use IMAP or leave pop3 mail on the server. I use IMAP and read my mail on different platforms and have it ready if I try a new distro.
If I happen to be in KDE and get my pop mail with KMail, why can’t I see it when later I open up Evolution (or Balsa or whatever…) . Why not have a standard location/format for mail?
Yeah ! I end up with the same conclusion as you, but I found a solution :
I pop my mail with fetchmail in my crontab, this go directly
to a local imap server on my computer, from which I can acces my mail by Evolution, Balsa, Thunderbird,…
I know that this solution will afraid the MS Windows Compatible User, but I love it 😉
The Office packages SEEM to be going that way (with KDE, Abiword and OO.org all migrating to using the same format)
Yes, I can confirm it. That is a great news.
Browsers. We have lots of pretty good browsers in Linux. Why do I have to set the same setting over and over for Mozilla, Konq, Opera, Galeon, whatever? Shouldn’t there be a common baseline of functionality (cache location and size, proxy info, Nescape plugins etc)
The freedesktop guys agreed on a standard hierarchy for storing user’s data, and are discussing a common format for storing bookmark (Xbel probably).
The first rule of HID: there’s no such thing as user error in operating something. If you don’t understand that, I don’t think you’re qualified to whine about the GNOME HIG.
-Erwos
If you take this to its logical conclusion, then isn’t the implication that if you have to learn something to be able to use it then it’s flawed? I’m not sure I can completely agree with that. I mean, if I give someone a copy of AutoCAD and they don’t know the first thing about architectural design, modelling, etc., then it isn’t a fault of the program if the user has no idea how to do things with it.
… when it comes to “computer interface”. So the article is generally useless as its premier question :
“…..Is there a single house that’s right for everyone? Is there a single car that’s right for everyone? Why is there supposed to be a single computer interface that’s right for everyone?….”
is meaningless.
regards
Suppose we only had a single desktop and it was based on KDE/QT. Now any commercial developer would be forced to pay royalties to Trolltech. Linux is about choice and it is wrong to grant Trolltech a monopoly like that. It’s not about the absolute amount of money, it’s the lack of choice and the creation of a monopoly that’s the problem. Nothing stops Trolltech from switching to $150,000 a seat ten years from now. Because of this some commerical vendors keep choosing and supporting GNOME.
If QT was LGPL instead of GPL Gnome might just die on it’s own. Without the LGPL change we are going to have two desktops forever.
that is the title of Linus Torvalds book why he created linux. I just started to code a little bit on a kde application. I have no clue about user interface design theory or whatever you like to call it. But I have some common sense and others have too. I am just coding for fun, if I ever release my code and someone likes it, it is fine. If someone thinks the UI sucks, then this is fine, too. I am not getting paid and so I am coding just for fun. Many other people are also coding just for fun. And I think it is great that people code for fun and give the code to other people. I’ve never had any problems with KDE and I like the powerful desktop it provides. If someone is not capable of working with KDE, maybe the problem is not KDE? For cars, you also have to learn how to drive. Computers and computer programs are much more complex, so why do people think they can use computers without knowing anything? Make’s no sense to me and even though one can probably use some programs without knowing how they work, because the UI is so simple, that’s not good for productivity anyway.
“If you take this to its logical conclusion, then isn’t the implication that if you have to learn something to be able to use it then it’s flawed? I’m not sure I can completely agree with that. I mean, if I give someone a copy of AutoCAD and they don’t know the first thing about architectural design, modelling, etc., then it isn’t a fault of the program if the user has no idea how to do things with it. ”
No one’s arguing that the user isn’t going to need to know the subject matter when using a highly complicated program like SAS or AutoCAD. However, if the user finds it difficult to pick up and use these programs with knowledge of the subject matter, the programmer is ultimately at fault.
Let’s say I give you an input field that’s only supposed to be given numbers, and labelled as such. Simple. Now, you type in a letter with your numbers. You hit enter, and the program beeps at you, or crashes, or otherwise does nasty things.
Who is at fault? The average poster I’ve seen would blame the user, especially if there was a little dialogue box that warned _not_ to enter. But, really, why am I letting you even type in letters when all the valid data is numbers? That’s programmer error! The user did something they shouldn’t have done, but the program should have been intelligent enough to _prevent it from happening in the first place_. It should have not entered a letter, and displayed a message (not dialogue box!) noting that letters ain’t allowed.
Like I said, HID is a complicated subject, and lots of unfun work to implement correctly. However, you and I both have experienced programs that were a joy to use, and ones which absolutely made you furious. A proper set of HIG is what makes a program a joy to use for the 5-95% quantile it’s aimed at, and usable for the rest.
-Erwos
I couldn´t say better. All this GNOME HIG blurb is getting annoying. There is only one gnome-centric distro and that´s RedHat, which is by it own claims a workstation/server purposes distro. And they do it to pleases people who _might_ be interested in developing something using GTK without licenses fees.
The distros with the ease-to-use emphasis like Lycoris, Lindows, Xandros, Knoppix and others have chosen KDE and there is a good reason. It isn´t just eye-candy. To most people used to Windows, KDE just seems more usable. At least, it _is_ for me.
I do agree that the lack of options must be appealing to some people, but not for everyone. I hate when I want to change something trivial in GNOME and have to open the GConf editor.
Instead of take off the user options, it would be better if those options would be better laid out in the preferences windows of the applications. There is room to both DEs to grow in this matter, not KDE only.
Regards,
DeadFish Man
Can you give me some examples of things you have to go into GConf to change? I’ve been using Gnome2 for a while now and haven’t felt the need to do that. It would help me understand your point of view if I had some examples.
BTW, Gnome2’s main problem for me is its immaturity. 2.0 did not feel finished (~pre-Alpha). 2.2 was almost finished (~Beta). Hopefully 2.4 will get even closer to feeling finished. Still, Gnome depends on a few long-term projects (like GStreamer), so it may be a while before it really feels finished.
I might be wrong, but didn’t you need to use GConf to change the UI colors, or to turn off opaque window moving?
And aren’t you tired of things like “to turn off ICQ autorun, you need to remove HKEY_LOCAL_MACHINESfotwareMSWindowsCurrentVersionRunOnce. Removal of HKEY_CURRENT_USERSfotwareMSWindowsCurrentVersionRun” will only work until somebody else logs in.”
For the rest, I don’t really see much of a problem. There are still things that could be done (common font sizes between KDE/Gnome, making copy-paste work between Gimp and OpenOffice) but for the rest most things seem to work good enough. Nobody ever says “Go use Internet Explorer because the Mozilla UI doesn’t use the Windows UI”. Nobody says “Don’t use WinAMP! It’s skinnable!” So a little difference in UI is not a problem, as long as they work roughly the same (and as long as I don’t need a magnifier to read Gaim from KDE)
And those confusing KDE options, can’t they easily be hidden between an Advanced-button?
I second daan. It seems much of the hatred of GCOnf comes from people who don’t use GNOME. GConf is meant to be a single aggregated way to control options. The GConf editor is NOT meant to be used for stuff like configuring the desktop. There is a big warning when you actually run it. The applications that use preferences in GConf are supposed to provide functionality to change the prefs, if the author wishes so. You can change those preferences using Gconf-editor, but that is not a solution for beginners. Beginners should stick with defaults anyway. Not that I do not like beginners, I was one once and I stuck with defaults then, but because if they do not understand what they are doing, then they should not do it.
However, I have used Windows for years, and I do not understand the registry, but I think I understand GConf quite a bit already.
In the old days you could (and still can for some apps) define your fonts in your $HOME/.Xdefaults. Too bad GTK/Gnome / Qt/KDE had to reinvent the wheel.
Making a gui for .Xdefaults would have worked. Hopefully freedesktop will come with a solution for this. Nobody likes to edit .gtkrc etc…
KDE has been successful because it provides the environment that the majority of linux users want, options, options, options, a few more options, options for those options, and eye candy. Gnome is boring, IMO. It’s a kiosk. Give me KDE any day of the week and I’m a happy girl. It’s hypnotic, I can get lost in it for hours and enjoy every minute of it. Infinite configurability, infinite posibilities. That’s KDE.
This issue was mentioned recently in the Gnome mailing list. I seem to remember Havoc Pennington (a leading Gnome developer) suggesting that the freedesktop.org site be overhauled; in essence becoming a hub for everything on the Linux desktop. Though I believe, the main goals of the project is to harmonize the underlying technology used in both KDE and gnome (D-Bus as an example), the first task is the synchronisation of release dates.
We usability folks are not saying that people shouldn’t be required to learn how to drive a car, as all new cognitive tasks require some amount of human learning; we’re just saying they people shouldn’t have to learn that the brake pedal is in the glove compartment.
Now, if some Junkyard Warrior wants to make some really messed up car with the brake pedal in the glove compartment (or even in the trunk), more power to him. He’s doing it just for himself. And if he wants to run this most badly designed vehicle certain to cause the deadliest of wrecks on his own 100 hectares of private farmland, that’s his perogative. If he makes his ludicrous design free and open to the public, then he certainly deserves a shiny gold star and a lollipop from Richard Stallman.
But I start to have problems when this very same Junkyard Warrior starts proclaiming that his car is “perfectly ready for the highway”, lobbies the public sector to try to get it put on every single public bus, tells people’s grandmothers the vehicles are perfectly safe to drive, tries to push his deadly contraption on ignorant CEO’s of trucking companies who don’t know anything about driving a truck and are dumb enough to buy into the ‘zero initial cost’ argument, and then when anyone has a problem, he proclaims “don’t blame me, I’m just a volunteer” or “quit whining about what you get for free”.
But I start to have problems when this very same Junkyard Warrior starts proclaiming that his car is “perfectly ready for the highway”, lobbies the public sector to try to get it put on every single public bus, tells people’s grandmothers the vehicles are perfectly safe to drive, tries to push his deadly contraption on ignorant CEO’s of trucking companies who don’t know anything about driving a truck and are dumb enough to buy into the ‘zero initial cost’ argument, and then when anyone has a problem, he proclaims “don’t blame me, I’m just a volunteer” or “quit whining about what you get for free”.
The problem with your arguement is two-fold really.
1) Developers of the messed up car does not proclaim their car is ready for anything. Distros push this linux on the desktop stuff and so do tech pundits and linux evangelists. The developers have not lobbied to get linux on public anything. The companies primarily that make money off linux push this. Even many of the largest projects are mainly supported by people doing the development for free because it is something they love. I have no clue why they bother. They get nothing but abuse from people who think because company A or B says that their distro is desktop-ready it gives them a free license to jump on every mailing list and tell them how their app, service or tool sucks.
2) Who is putting the brake pedal in the glove compartment? Following your own analogy car makers have consistently toyed with the “interface” from gear shifts on the steering column to ergonomic control designs for the radio. The interface differences between Gnome, KDE, Windows XP and Mac OSX are much closer to these kinds of differences as opposed to the type of wild example that you are putting out. Linux may not be desktop-ready but it is not some dangerous killing machine ready to take your life. Hyperbole thy name is Ilan in this case.
Personally, I wish the whole general linux desktop crap would die. If the distros focused on the core *Nix conversion market and marketed toward converting Unix users in the corporate realm to a linux solution they would be better focused and in better shape overall IMO. Think about the fact that companies like Exceed have made money off of Unix focused companies for years. Before the *Nix development staff in my office and the system folks moved to linux we spent 90% of our time in an Exceed session. It made sense for our business.
My Name is Joe I-Want-to-Configure-My-Car-to-My-Heart’s-Desire.
I think it is good for me to be able to have a moving gearshift, you know, one that allows me to move its position from the centre console, to between my legs, to the door, or behind the steering wheel.
Why because, maybe,
1. I like driving with my hands in my lap, so the gearshift might as well be right there too.
2. Or, I like talking on the phone whilst driving, so I want the gearshift on the door so I can operate it with my driving hand.
This is maybe more like it. Its silly I know, but there is nothing wrong with enforcing rules not because they make thing s necessarily better, but that things become more consistent.
I like to program software and distribute them non-commercially, but I dislike sharing my code. Why should I have to pay for QT if I want to distribute my programs for free without code? That’s why I would choose http://www.wxwindows.org if I wanted to code for Linux. If I later want to make a commercial product, I will probably do this with wxWindows, because that’s what I’m used to. One less potential customer for Trolltech.
If you look at Windows freeware, you will find lots of high quality software without source.
“but for now, i’m using KDE. i finally got my wire-frame mode back, i’m tired of opaque windows being forced down my throat.”
I can understand your frustration there. I agree that there are performance issues with metacity and gtk2 just in terms of draw performance. I do believe the situation will improve in time, and 2.4 certainly seems to be a bit snappier. However, you might be pleased to know that there is a “reduced resources” mode in the 2.4 release candidate which, among other things, does provide a wireframe drag mode. It’s not pretty but it works. I believe the key is called reduced_resources, under apps/metacity. I managed to port Sun’s wireframe patch (which is more aesthetically pleasing, IMHO) to an older version of metacity (for Gnome 2.2) but it messed up the keybindings and I’m not skilled enough to resolve it.
If you happen to use Linux, the other thing is that I noticed a big performance boost with X in general using the WOLK kernel which is far and away my favourite 2.4 branch. In fact, although certain operations have proved to be more speedy with 2.6 I _still_ get the best graphical performance with WOLK. Might be worth a shot, even if you stick with KDE. The q->full low latency patch (courtesy of Chris Mason) seems to work nicely for desktop performance too (run elvtune -b1 on your block devices).
but there is nothing wrong with enforcing rules not because they make thing s necessarily better, but that things become more consistent.
Sure, there isn’t anything wrong with that.
That way the anti-gnome trolls can come out of the woodwork and complain constantly about stripping configurability options out from the desktop.
I imagine there are ways to “lock down” the desktop for KDE. (There is such a thing for gnome.) That way users don’t customize everything to the point that the admin jumping on the box does not have to take an extra second to think about where to find things when they come to fix your setup. I know this is the reason that Windows admins hate litestep and Windowblinds and all that stuff right?
Most of the defaults for all desktop environments including the commercial ones like XP and Mac OS are not that hard to pick up even for a relative neophile.
In fact, give me a neophile over some OS specific hotshot who wants to complain about how their PC does not act like their Mac at home or how the Mac at school does not act like the Dell in their basement blah..blah..blah.