Slashdot Mirror
KDE 4.1 Beta 2 – Two Steps Forward, One Step Back?
jammag writes "Linux pundit Bruce Byfield takes a look at the latest KDE beta and finds it wanting: 'Very likely, KDE users will have to wait for another release or two beyond 4.1 before the new version of KDE matches the features of earlier ones, especially in customization.' He notes that the second beta is still prone to unexplained crashes, and goes so far as to say, 'Everyone agrees now that KDE 4.0 was a mistake.' I'm not too sure about that — really, 'everyone?'"
"Everyone" agrees that Vista is "a failure", even though it's really not. So why can't dumb generalizations be applied to software that's supposed to be perfect in every way?
You know, I thought that the idea of Beta software was so that people could report unexplained crashes back to the developers....
I'm starting to think GNU is the problem with "GNU/Linux" these days.
I certainly don't think KDE 4.0 was a mistake beyond calling it "4.0" which led a bunch of idiots to expect something "finished", and that despite the up-front warnings that it wasn't finished.
It's a clear design improvement on 3.x in every way (though I don't particularly like or use the new desktop with its "plasmoids", I didn't like the 3.x desktop either, and the 4.x desktop can emulate it trivially - desktops widgets are just pointless, you just don't see them or the desktop for 99.9% of the time you're using the computer), it's just not stable yet.
One of reasons may be that they are doing it for free, in their spare time. Not eight hours a day, with their paycheck dependent on the quality of the result and with best professional artists, designers, usability specialists etc hired for big $$$ to decide what is best.
As much as we want to think otherwise, most of open source software is amateur production. Some of it is professional in means of program, but great most is amateur when it comes to UI design, art, and such.
Neither is all that happy, both have been looking forward to a fully usable kubuntu with the 4.1 (because it "seems more like windows"), but maybe I should begin looking into E17 for them?
Or perhaps they can stop expecting it to be something it isn't and get used to Linux as a real operating system, not "that shoddy free Windows clone" they expect it to be.
If there's anyone I hate more than stupid people, it's intellectuals.
It probably means just that. Familiar paradigms are important to a lot of people.
blah
Here's why:
1) NIMBY - If Z is a feature or program I don't use, not only do I not care about it, I don't care about whether or not it can interact properly with programs I do care about.
2) Windows-ism - Many projects now try to replicate the functions of Windows apps. But the clones and work-alikes they produce are not only imperfect, programmers also can't take the same shortcuts that the Windows developers do.
3) Real Programmers - If a program isn't hard to write, it isn't worth writing, and if you make it easy for programmers to write for a platform, especially new ones, they will only produce crap that you somehow have to deal with. Compare this with MS's "Developers developer developers" motto, or Apple's excellent dev tools.
4) Esoterism - The command line is better than graphics. Graphics, and especially graphic quality is unimportant, and studies with evidence to the contrary notwithstanding, whether an interface is cleaner or more obvious or better-looking is irrelevant. It's okay for GUI tools and programs to just be front-ends for their command-line equivalents, even if it puts unnecessary limits on the graphical version.
5) Arrogance - (related to 1) There is only one right way to do things, one language, one library, one kernel, one package, one work-flow set-up. If you do it any other way, you're wrong; if you suggest that another way is good, I must shoot you down and insult you because you implicitly threaten the validity of my worldview; if you say that there can be more than one solution to a problem, you are really saying that your solution is right and mine is wrong.
I once listened in on a conversation by some digital typographers about their work set-ups, and unlike linux-heads they were genuinely interested in the advantages and disadvantages of different ways of solving the same problem, instead of arguing over whether which was best.
Short answer? Mnemonics.
Long version: it's easier for most people to fudge through something they vaguely remember doing by pictures than it is for them to memorize a set of arcane terminal they vaguely remember. People who do things other than program and learn Linux inside and out have all sorts of other random esoteric knowledge buried away, and there's only so much that a single person can keep in their head. These people are called end users, and frankly, if you don't understand why politely asking them to "simply" learn the terminal commands is a mind-numbingly stupid proposition, I seriously recommend staying the hell away from UI design.
Try not to take me more seriously than I take myself.
4 is almost a complete rewrite. It seems people have the impression that the reason all of the 3.5 desktop features weren't completed in 4.1 is because of a conscious choice. When actually, it is was just limited time. Feature freeze tends to stop the adding of magic ponys.
KDE shot itself in the foot by making the KDE 3.x so polished. KDE 3.5 is essentially 9 years of evolutionary development from KDE 1.0. Unfortunately, its impossible to recreate 9 years of development and polish in only 3. I think that the long term prospects for KDE 4.x are great, but short term I'll continue to use 3.5.
I've tried the first beta of 4.1 and while its much more functional than 4.0, its still not there and probably won't be for a few more releases. On the other hand, I remember that KDE 3.0 was, while more functional than 4.0, also much rougher than 3.5, so I can't complain too much.
Why?
Aren't you aware of the Linux development paradigm that has been the rule since Linus released Linux? ... "Release Early, release often!" FOSS depends upon the users helping in the development of software, not whining about perceived or real problems.
Bruce Byfield summarized his findings with the following statement:
How stable KDE 4.1 will be when released at the end of this month is anybody's guess. But, judging from its features, the release will be a major milestone in the 4.x series. Unfortunately, it will almost certainly not be the complete answer to user discontent that has been promised. It might even drive large number of users away from KDE altogether.
Such a reaction would be misguided. KDE 4.x has many features, including the use of scalable vector graphics and natural language searches that make it the most innovative free desktop currently in development. Moreover, if you dislike some of its experiments, you can work around them with no more trouble than it takes to change your desktop wallpaper -- for instance, one of the widgets you can add to the desktop is a KDE 3.5.x menu.
That is wise advice.
Troy Unrau introduced KDE4, before the first beta was released, on Jan 1, 2007 with the "Road to KDE4" series at http://dot.kde.org/1167723426
Before he resigned KDE4 to focus on his Masters in Geology degree, Troy posted the following comments:
We knew there would be some pushback to the major changes in KDE 4.0, because, believe it or not, history is simply repeating itself. KDE 2.0 was met almost exactly the same way, although open source was flying a lot lower under the public radar in those days. It took until KDE 2.2 before distros mostly stopped shipping KDE 1.1.2 and were happy with 2.x.ferent standard. Somehow though, there's still a lot of positive press about KDE out there, which means that the developers have done something right (or us Marketing guys are worth our weight in Rhodium...) and the naysayers have not killed a project they confess to love.
So my message to all the disgruntled users out there are: use KDE 3.5.x, and wait until 4.x makes you happy, or better yet, help. That's what the Mac OS users did. That's what the Apache users did. That's what our KDE 2.x users did. The software you are getting from the KDE project is free, worked on by a team of developers that actually like to use their own software. Improvements are coming fast, and KDE 4.1.0 is scheduled for July. 4.2.0 for January, etc. If you use 4.0.x, have found issues, and would like to help improve 4.1 before the release, grab the SVN version, using KDE4Daily (virtual machine image), the automated kdesvn-build script, anonsvn, and file bugs. Join the bug squashing days that are announced via planetkde or the dot. And bring a positive attitude because KDE is yours, just as much as any coder!
The hysteria in some complaints (and deliberate FUDing and astroturfing in others) is misplaced. FOSS software is not static, especially when there is a vibrant body of users CONTRIBUTING to its development (coding, testing, documenting or donations). Users who do not contribute but only complain are "poisonous users". A project grows when it has an amply supply of contributing users. Any project dies when its users are poisonous.
It is also obvious that some "complainers" are not KDE users at all. Their motives are obvious. A lot of this brouhaha has been exploited by a few bloggers trying to increase their page hits by inflammatory comments with little basis in fact.
Running with Linux for over 20 years!
It's okay for GUI tools and programs to just be front-ends for their command-line equivalents, even if it puts unnecessary limits on the graphical version.
On the other hand, there's a pretty strong argument this should always be the case EXCEPT for the tools that build the GUIs themselves.
Consider the standard menu of a program[1] where you'll find the same options from the File menu almost always as buttons in the application right under the file menu and you'll find the edit menu items in the context menu.
Point is, there are plenty of ways to display these UI options to the user. They can and should be separated from their actual implementations. This would ultimately mean that the UI can be generated according to a user's personal preferences and needs (including assistive technologies or device limitations) while the actual guts of the application stays the same.
At least, I believe this is the way forward for GUIs.
[1]
[...]
-HobophobE
Nothing laughs forever.
The full context in which I call KDE 4 a mistake is: "Everyone agrees now that KDE 4.0 was a mistake. However, what the mistake was -- and whose -- is a matter of opinion. KDE developers blame distributions for rushing to include a release that was never intended for everyday use, while users blame developers for changing everything." In other words, all I'm saying is that it's causing a lot of controversy
Err.. no. That is not at all what you are saying. If everyone is in agreement on a point, there can be no controversy on that point. Anyway, the "everyone agrees ... " statement was the most interesting part of your article and I was displeased to see that it was just grabbed out of the blue. If the KDE core devs feel that 4.0 was a mistake, getting to know why, what they think they should have done different and what they have learned would have been very valuable to know for other developers. If distributors feel that distributing 4.0 was a mistake, then I would like to know what they will do about it? Will they be more strict about upgrading to flaky libraries?
But it is extremely uncommon for developers to admit that they have made a mistake. And I very much doubt that the KDE 4.0 guys think it was a mistake. You definitely made a mistake if you thought that an "everyone agrees" statement would slip. :)
Football Odds
And so does any GUI.. they all have icons, some sort of "OK" buttons, a close button, etc.
KDE 4 is probably more different then Windows then Gnome. Just because Gnome's main "bar" is on the top, doesn't make it somehow completely different than Windows. Move the bar to the bottom, and BAM, you have a Windows-looking UI.
- It's not the Macs I hate. It's Digg users. -
KDE 4.0 and 4.1 are not meant to be perfect in every way. They are meant to establish a new scheme of APIs and a new design dynamic. It is a big overhaul that is in its beginnings. Nobody is claiming KDE 4.x is feature comparable to 3.x right now. This is just one person's view, and this is another view with excellent counterpoints. It is a failure where people are expecting too much of it in its current state.
Vista is supposed to be a workstation solution ready for every day production use right now. People are considering that to be a failure in its current state as well, and you are right, these two alleged failures are similar. But one product that is at an early start (4.0 & 4.1 beta, the more mature 3.5+ still seeing a lot of active development and use due to its maturity) and the other has the promise to be mature enough to use right now. You are not forced to upgrade to KDE 4.x, but Vista is required for some of today's games and applications because they don't run in earlier versions. This is the difference.
Twinstiq, game news
Vista failed to achieve the goal that MS had when designing and programming it. I'm not sure how that can be anything other than a failure. The fact that they're really having to pull the plug to get people to move on and that people will likely switch directly to Win 7 if they can will prove it. And I see no evidence that that's not going to happen.
As for KDE, make it less bloated, better modularized and make the defaults include fewer programs.
I stopped using Windows because of the bloat and the unwanted features, I'm not about to start using a desktop environment that's as bad. But, really the same could be said for gnome and pretty much every desktop environment.
And for the love of god allow some alternate way of compiling the smaller applications without KDE itself. I hate having to install both the gnome and KDE libs because there's that one program which invariably requires the other set of libraries.
KDE? more modular?
Kparts means that you can include practically entire programs (spreadsheets, browsers, editors) inside other programs - how much more modular can it get?