Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
KDE GUI

Criticisms of KDE 3 Release Process 201

Posted by michael from the few-raindrops-must-fall dept.
An anonymous submitter sent in a link to a recent email from the kde-devel list, criticizing the release process. Hopefully the KDE guys can work out any problems and keep up the good work that we've seen in the past. Update: 03/10 14:20 GMT by M : One of the comments below points out that another KDE developer has made an extensive response to the original criticisms.
This discussion has been archived. No new comments can be posted.

Criticisms of KDE 3 Release Process More

Criticisms of KDE 3 Release Process

Comments Filter:

  • Take it Easy on KDE! (Score:3, Insightful)

    by guinan ( 191856 ) on Saturday March 09, 2002 @08:52PM (#3136435)

    After all of the people recently complaining about the possiblity that previously open sourced software might have to succumb to the relentless tide of capitalism, I find it somewhat confusing that they are willing to find fault with KDE.

    Right now, KDE looks to be the best hope for Linux to enter mainstream as a desktop OS. A VERY major portion of the impetus for users to not change is a lack of familiarity with the desktop. They don't care how it works, or what it does. They need their computer screen to be familiar to them the first time they experience their new OS. Once they're comfortable with it, they might abandon the "Windows look", but until then it will get users.

    KDE could probably use not criticsism, but instead help from able computer scientists who want to see the open source movement triumph
    • The mail was a rant from 3 core-developers who don't feel compfortable about the changes to the "release-policies", made for the 3.0-cycle. It sure helps to see the disadvantages of a procedure to optimze it. That's, what is happening in the thread on the list right now and it's not even worth a slashdot-message, because it's absolutely no big deal. KDE will not suffer from it, quite the contrary, work-processes gain much, if the workers are capable of constructive criticizm.
    • KDE != Linux. KDE can stand on its own, can Linux? I used to think so, but so many people beg to differ that I'm not sure any more.

      • KDE != Linux. KDE can stand on its own, can Linux? I used to think so, but so many people beg to differ that I'm not sure any more.

        I tend to think so, although KDE is an excellent choice for a windows manager, it is not the only game in town, beyond even Gnome. I do not use either KDE or Gnome as my desktop, I use ROX, so I could uninstall both and 90% of my apps would work fine. The issue is not now technical nor has it ever been. Instead it is an issue of user preferance. Certainly loosing either KDE or Gnome would be a hard blow, but I think Linux could survive loosing both. Though I think the chance of loosing even one is nonexistent let alone both.

    • "A VERY major portion of the impetus for users to not change is a lack of familiarity with the desktop..."

      I just wanted to mention that it wasn't the interface so much that makes me not us Linux, it's that I have no clue how to get hardware to work. Windows has me that spoiled. I'm so used to having neat little menu driven things I can run that allow me to get things up and running, that when I went to use Linux (KDE on Redhat, I believe), I was unable to find what I needed to get network and sound going.

      You may chalk this up to me being a helpless newbie, but I don't have this problem with Windows or even Mac. I guess what I'm really saying is that KDE at 2.0 was fine, but tinkering with the hardware to make everything work was what turned me off. I admit that I didn't put all that much effort into it, but I only have so much time, you know?

      Apple got the idea right, though. Look at what they did with OSX. They built upon the BSD Kernel (I think it was BSD... don't shoot me if I got it wrong. Please feel free to correct me, though) and made the interface with a target audience in mind. The result? I have a coworker who is able to tinker with his Mac, but he's never needed to know the root password to his machine.

      Redhat's gotten close to this in 7.0. I really feel like 8 or 9 may be enough to get me going in the Linux world. A new version of KDE is icing on the cake.

      • Try SuSE - all the sound, display, printing, scanner, networking, etc. settings are nice and neat in KDE's Control Center, along with all your font, email identity and wallpaper settings. Everything is nice and centralized.

        SuSE is really the nicest desktop OS I've used, IMO, possibly barring OSX. (That's *just* talking about the desktop experience, and not apps or OS). It blows away Red Hat and has a leg up on Mandrake... I'm looking forward to their KDE 3.1 based setup.

        Incidently, so many features are hanging on the feature freeze for KDE 3.1 that that is the release that I'm really looking forward to. 3.0 is an updated API and updates on the core apps with very few killer features, other than speed and core changes... 3.1 is when the nifty features (many of which are already written) get integrated. Several of the developers who have pretty much wrapped up their code for 3.0 are seriously looking at 3.1. Remember - 3.0 is primarily a port and rework to provide a faster stable core with more functionality. 3.1 will actually use this new functionality to add new features.

        --
        Evan

        • nice and centralized

          Yuch. Different strokes for different folks indeed.

  • Hello! (Score:4, Interesting)

    by brandonsr ( 550431 ) on Saturday March 09, 2002 @08:56PM (#3136451) Homepage
    KDE3 is early beta software.. Take a look at his first comments

    1) Packages missing from the release entirely (1)
    2) Rampant compile problems
    3) Last-minute changes to build requirements that cannot be met by
    many developers without an operating system upgrade (2)
    4) Many outstanding bugs (3)

    of course there's going to be problems with the software.

    no one has made a post about the development of Woody and there are a lot more problems in it than KDE3. isn't this all part of the sotware development proccess? (emphasis on development)
    • >> 2) Rampant compile problems

      i don't think you should be checking in code that doesn't compile...
      • That's true. But I have compiled the KDE3 code and it worked fine. I think for that he's reffering to the cvs code. I've compiled a few things vis cvs and had one or two compilation problems. But that's to be expected, or so I thought.
      • It DOES compile. The issue, the author was refering to, is already sorted out. Don't worry, the mail is just the usual occurence of panic close to releases.
    • KDE 3 is early beta software? I agree, however it shouldn't be labeled as a release candidate, and therein is the problem. An RC is an almost final product that if no more bugs are found, will ship as is. KDE 3 doesn't even compile at the moment, and doing a hack job to get it to compile results in an almost unusable system. This is what his beef is with the KDE 3 release process, and his claims are fully substantiated.
      • I have just compiled the latest cvs and haven't had any problems. Actually, I'm amazed at how many bugs have been fixed since beta2. I still haven't encountered a single bug yet. It's faster, more stable, less buggy.

        This is on woody, and the only thing I've had to upgrade was Qt (from 3.0.1 to 3.0.2)

    • Re:Hello! (Score:5, Insightful)

      by Anonymous Coward on Saturday March 09, 2002 @10:23PM (#3136651)
      1) Packages missing from the release entirely

      Release Candidate != Release

      2) Rampant compile problems

      Really? Compiles fine here and elsewhere. Of
      course, if it doesn't compile for you the proper thing
      to do is report the specific problem to
      the mailing lists immediately.

      3) Last-minute changes to build requirements that cannot be met by
      many developers without an operating system upgrade

      Hehe, this one was funny. He was complaining
      about a requirement for developers compiling
      out of cvs       to upgrade to gnu autoconf > 1.50.
      Those compiling from the release tarballs won't
      even be affected, and the "operating system upgrade"
      consists of downloading autoconf and
      compiling it. Took me literally 3 minutes.

      4) Many outstanding bugs

      Well sure, all software has bugs, even when released.
      Released software shouldn't of course have
      critical bugs so when you find them, you
      should report the specific problem to
      the appropriate forum (mailing lists at this late
      date in the cycle) immediately.

      This whole thing was just an overation from 3
      developers who felt left out when they weren't
      invited to the recent KDE hacking session. It's
      unfortuate, but the sky is not falling.

      The best way to help KDE is not complain but
      to download the release candidates, compile
      them (yeah, you can do it - ./configure; make;
      make install), test, and report bugs.
      • On my machine, by the time I finished compiling it, it would be released. Don't discriminate against those of us with 400MHz and 48MB RAM!
    • Last-minute changes to build requirements that cannot be met by many developers without an operating system upgrade
      What does it mean? newer kernel?
    • "While Dirk Mueller is respected throughout
      the community [...] New leadership for KDE 3.1 is needed. I call for a vote of no-confidence in Senator Mueller's leadership."

  • Professionalism == Bad (Score:5, Insightful)

    by lkaos ( 187507 ) <anthony AT codemonkey DOT ws> on Saturday March 09, 2002 @08:57PM (#3136461) Homepage Journal
    The best part of developing free software is that it is low stress. People tend to get all bent out of shape about it. I think this is a pretty good example of what happens when people get stressed out about something that people are for the most part doing because they enjoy it.

    So the KDE guys got together and were inspired to perform lots of last minute hacking. More power to them! So what if the 3.0 release is delayed or has a few issues. I think these three guys who signed the letter were just jealous because they weren't involved in the process.

    I don't use KDE, and never liked it, but I have to stand up for the developers here. Just enjoy developing the software and stop bitching because there aren't 'hard freezes' before a release.

    • Re:Professionalism == Good (Score:4, Insightful)

      by Arandir ( 19206 ) on Saturday March 09, 2002 @09:36PM (#3136552) Homepage Journal
      Professionalism is a Good Thing(tm). However what your PHP calls professionalism might just be artificial. Sticking to a release date no matter the state of the code is unprofessional.
      • Sticking to a release date no matter the state of the code is unprofessional.

        True, but if you don't do that, there will always be a few developers saying the code is not ready, and there's still work to be done. That's how releases end up getting ridiculously delayed.

        It comes down to the responsibility of the developers. If there's a freeze coming up, have enough common sense to know that it's not arbitrary, and it's for the good of the project. Don't barf all over the code 2 weeks ahead of time just to get a whiz-bang feature in.

        • After a feature freeze the only thing that goes into the code is bug fixes. Never, never, never put in features after a freeze. But always allow in fixes for the highest priority bugs.

          It helps if you have three freeze points, one to stop features, one to stop minor bugs, and one to stop everything but the mandatory bugs. You also need a release manager with very thick skin armed with a spiked club.
      • Well, if you want to talk about professionalism, it also is very unprofessional to bad mouth a project right before it's release.

        A lot of the time, developers like to blow things a bit out of proportion and that's all well and good when your just around developers, but it is very bad to make such public comments.

        We always had a general rule of thumb that we follow. Within a month of a release of our project, we never use the word 'core dump' around the management. Dumps are no big deal most of the time but it doesn't give management a warm and fuzzy feeling to hear the words 'core dump' so close to deadline.
    • You don't like KDE? Maybe if you used WindowMaker you would!

      Before I saw the light (WindowMaker/slackware), in the very beginning, I used Gnome. I liked it. I mean, after all, KDE _does_ have that windows wannabe look about it. Also there's that whole QT thingy.

      Then I used WindowMaker. I still have kdelibs and all, just because the distro came with it. I use their nifty kmahjong and my brother still loves kspaceduel. And kword once in a while too. Do you know how much better kde apps look, compared to gnome apps? In windowmaker, (and NeXT, OS X) we have this thing called AppIcon. Some bad acting apps (like SDL programs and XEvil and xpdf just off top of mind) doesn't make AppIcon, so I have to emulate it. Most all of x apps just have AppIcon with (ugly) default icon. Some x apps (_all_ of kde programs included, also staroffice and gv from top of my mind) even supply it's own icon so that I don't have to hunt through /opt/gnome/share/pixmaps to give it the correct icon. No gnome apps have ever done that. All kde programs have their own icon for AppIcon. It just makes my life more beautiful and easier.

      My bro covets my linux installation and wants to try linux out (maybe this summer). When he does, I would recommend KDE, just because it would make later transition easier. (He really likes windowmaker too.)

      I don't know anything about X11 programming, so I don't know what the nature of these AppIcon is, but I know that they are a good thing. I may not like the whole K _desktop_ environment, but KDE apps surely are a Good Thing!

      • Also there's that whole QT thingy.

        What exactly is the "thingy" about Qt? Some of us happen to think that Qt is a very professional, elegantly designed toolkit. It's been GPL'd for something like 1.5 years, so FSF zealotry shouldn't be an issue either...

        Seriously, I have a hard time figuring out why so many people bash Qt. Maybe someone can enlighten me.

        • jeez, "there's" could be "there was" too, you know. I have nothing against QT. I'm just mentioning it because there _was_ a "QT thingy." I also happen to know that it's fully resolved. I just say it because when I first started using linux, I asked my linux buddy about kde and gnome and which to use (cus I couldn't decide). He told be "use gnoem cus it's more free." So I did. Maybe I should have included the whole explanation. Probably not.

          If you didn't get my post, it was pro-KDE. In fact, I type this on Opera which uses QT.
          • Okay, "there's" -> "there was" clarifies things for me.

            Sorry if my previous post came off a bit flamish--that wasn't my intention. I am simply curious why Qt continues to have a bad reputation amongst many slashdotters. I guess I was asking the wrong person.
        • Seriously, I have a hard time figuring out why so many people bash Qt. Maybe someone can enlighten me.

          It was written in Norway, where they eat fish for breakfast, fish for lunch, and fish for dinner also. No good red-blooded American would have anything to do with such a fishy product. (Seriously) it's just ordinary American xenophobia.

          PS no, I'm not American... but then I understand irony and like Qt.

      • You don't like KDE? Maybe if you used WindowMaker you would!

        Bah! WindowMaker was too much a resource hog when I first installed so many years back... I could only open one window and then if I tried to open another, I would run out of memory. Those were the good ole days :)

        I'm not much a GUI person. I use gnome-terminal, Emacs (not XEmacs, but GNU Emacs), and mozilla and that's about it. The Gnome stuff seems more light weight to me and I like the fact that it works well with Sawfish (since I was able to do some lisp hacks to sawfish to add macros and some other neat stuff).

        Give me a few terminal windows, and a display to bring up emacs on, and I'm happy as hell.
        • Well, that certainly depends. I could run windowmaker on, say, pentium, and it would work great. Well, great until I start Mozilla! (or Opera for that matter, which is what I use. But Opera is less of a hog, it's pretty much the fastest GUI browser out there. but I digress.) My point is, with windowmaker, your program is the hog, _not_ the manager. With KDE or Gnome, the manager will certainly be a hog itself.

          The Gnome stuff seems more light weight to me

          I assume you don't mean Gnome environment vs WindowMaker.
          If you mean Gnome apps, I guess they could be lite. But that's not WindowMaker's problem. Just use gnome apps in WindowMaker.

          WindowMaker is known as a lite desktop itself while providing pretty cool environment. Sawfish w/o Gnome seems pretty spartan to me.
        • you really should try xclock -- it helps me get to bed in time for work.

  • Not Release Problems (Score:5, Insightful)

    by BadlandZ ( 1725 ) on Saturday March 09, 2002 @09:02PM (#3136476) Journal
    Look at the artical, it criticizes decisions made, but it's not criticizing KDE itself, or any release of KDE. It's critical of the _process_ not the _release_!

    It's a failing of leadership (if the criticism is true). I think it's important to remember 2 things here:

    • KDE releases are important to the acceptance of Linux on the desktop. More judgement of the FINAL PRODUCT should be focused on, not just some little pain in the ass details about getting it ready.
    • OK, thousands of programmers coding for you, for free can't be a bad thing, ever.

    In light of this lack of management discovery, maybe a couple programers will start to see all the recent criticizm's of software managers (as in recent stories [slashdot.org] here) may be not as useful as trying to actually support managers of projects (espically OSS ones) a little more.

  • This happens in large projects. (Score:5, Insightful)

    by JanneM ( 7445 ) on Saturday March 09, 2002 @09:02PM (#3136477) Homepage
    This is in no way unique to KDE; large development projects sometimes stumble even with the best of intentions by those involved. The Open Source community is unique in that everyones dirty laundry gets aired in public. This can make the process seem unruly, haphazard and chaotic compared to closed development - the truth is that the same kind of conflicts, friction and occasional disasters occur there as well, but hidden from view.

    I'm not a KDE user myself (I prefer gnome), but I'm confident and hopeful that the KDE development team will get past these problems and produce another good release. They've done very good releases in the past, and there's no reason for them not to do it again.

    /Janne

  • Whew... (Score:3, Funny)

    by Lobo_Louie ( 545789 ) on Saturday March 09, 2002 @09:05PM (#3136485)
    ... I'm glad Gnome is perfect.

    :)

  • In the early days of KDE, I had a heck of a time getting it to build on solaris. kdesupport was the worst offender because the sound support was linux-only. Over time they got smart and and made many non-portable routines optional.

    Maybe kde needs something like mozilla's tinderbox - then whoever breaks something in only some environments would be "on the hook" to fix it.

  • Make sure to get both sides of the story (Score:5, Informative)

    by klieber ( 124032 ) on Saturday March 09, 2002 @09:07PM (#3136492) Homepage
    The post linked to in the /. article is one side of the story. To get both (all) sides of the story, check out this thread:

    http://lists.kde.org/?t=101566017800001&r=1&w=2 [kde.org]

    And specifically, Dirk Mueller's response:

    http://lists.kde.org/?l=kde-core-devel&m=101567612 207504&w=2 [kde.org]

    I'm not saying one side is more right than the other -- merely that there are certainly two sides to this issue.

    • This is Neil blowing things out of proportion again. It's honestly nothing new. He has long history of lambasting anything to do with KDE3.

      Others are having *excellent* experiences with KDE3. Just check the "fucking amazing" (sic) post, for example.

      So yes, there are two sides to the story
      • Neil here,

        What post are you talking about?

        Now I'm going to have to start looking through your past posts to figure out just who you are. From the name I have to guess you're one of the dot.kde.org editors, but I'm not sure.
      • Neil did not blow anything out of proportion. Uncensored opinions like these are the best, because you need some unpopular opinions to evaluate your goals and methods and so you can improve them along the way.

        There have been some questionable events and changes in the past week within KDE development and the upcoming release. Since noone had yet the courage to express concerns about these changes, Neil stood up and did, with a firm voice.

        The result is some strong but constructive discussions to improve the situation and do what is best for KDE. A second RC has been created and the results of testing it will decided if there will be a final 3.0.0 release next or a delay if necessary.

        Yes, us KDE developers disagree sometimes. We're not always best friends. But we respect other opinions and keep our common friend in mind: KDE itself.

    • I don't buy it (Score:3, Informative)

      by dimator ( 71399 )
      I don't buy it (the complaint that is). The reasons for the changes seem to be for the long term good of KDE, and to keep the breakage of going from 3.0 to 3.1 in the future to a minimum.

      I have found the KDE guys' release scheduling and management of high quality in the past, and judging from the minimum of hiccups I got building 3.0 RC1, I can say they're still top notch.

    • merely that there are certainly two sides to this issue.

      Hallelujah! Now, could someone beat the Slashdot and Newsforge *kaf* "editors" with a cluebat so they'll figure that out?

      If they have to post crap like this, the least they could do is give a direct link to the actual mail archive and a link to some of the rebuttal.
  • KDE isn't the only open-source project to jump the gun. Maybe the should release a KDE-3_dont_use tarball ;)
    I personally don't use KDE because it feels cluttered and slow. I usually use Enlightenment 0.17 CVS (when I can compile it). It looks great, is lightening fast, and is already very stable.

    Anywho, whatever happened to Katabase, of KOffice? I have been waiting for that before I try a switch from OpenOffice to KOffice...

    • Re:Assorted things... (Score:1, Interesting)

      by Anonymous Coward

      From what I can gather, KWord and KSpread should be usable* beginning with their next official release. They won't be feature-competitive with OpenOffice until quite a bit later, but they'll be nice enough for regular use. Katabase has a long long LONG way to go before it is usable. I have no idea about the other KOffice apps because I haven't been following them. I frankly think KWord's frame-based DTP design is going to be the thing that puts it on top in the long run. Then again, OpenOffice works NOW, and there's something to be said for that...


      * my definition of "usable" may not be the same as yours. Same goes for most things.

  • ...nobody wants to give me a quality desktop environment for free...

  • New leadership for KDE 3.1 is needed.

    Beware the new Chancellor doesn't become and evil Emporer!

    • Neil here,

      Well, the vote of no confidence won't come on Dirk until after 3.0 is out, for better or worse. :-)

      Question is, which one of Ryan, Charles, and myself is Padme?
      • To add some perspective, Neil (and I) have both considered being release coordinator for upcoming releases. So we are definitely not just whining, we are prepared to do something about it. My only reason of doubt is whether I can find enough time to do a right job without compromising work and study efforts.
  • My shop develops web based database software for the Marine Corps in the Okinawa, Japan region. Every application we support is based on a core set of libraries to handle all the ugly parts of talking through ODBC and the mundane parts of HTML Tags.

    Although I've been coding for almost 16 years now, I've never been involved as a project manager for something this big. (our main app is rapidly approaching 100,000 lines) We have found through trial and error that code freeze/documentation periods are essential to ensure that we are all still using the same vesrions of the core libraries. This is especially critical for web page design as each page can almost be considered an object with a specific interface. If you change the interface on a page, you just broke every page that connects to that page. I'm sure the various components of KDE are no different.

    KDE dev team, don't shoot the messenger! I think this is a fantastic opportunity for you to have your development practices analyzed by the slashdot community. I don't even think you could hire a consulting firm this honest and experienced.
  • If they use the same 4pt font in the article that they use in their code, no wonder nothing gets fixed!!!

    You can't re-code what you can't read.

    .
  • <sarcasm mode=overfiend>
    I didn't realize Adrian Bunk had decided to "participate" in the KDE release process too...
    </sarcasm>
    • Come on, I know Adrian personally from real life.

      He did a _lot_ of work for Debian QA. He did maintain more packages than most (or perhaps any) other developers and was one of the No.1 Bug fixers. This guy here is just a troll.

      regards,

      Michael

  • On a related note (Score:2, Interesting)

    by IronDragon ( 74186 )
    My recent builds of kde3 are looking rather nice. Everything runs smooth, the fonts look good, and a lot of older bugs no longer show up.

    Whats the problem here? :)

  • word of reason from waldo bastian: (Score:5, Informative)

    by Karma Sucks ( 127136 ) on Saturday March 09, 2002 @10:01PM (#3136608)
    From: Waldo Bastian
    To: kde-core-devel@mail.kde.org
    Subject: Thoughts about releases.
    Date: Sat, 9 Mar 2002 15:47:53 -0800

    Releases are a funny thing, you know, we love them and we hate them. Before a
    release everyone tends to get all excited and some people see a bunch of bugs
    and problems and then go into panic-mode. I have seen it happening with about
    every KDE release. Psychology must be playing tricks on the human mind.
    Looking back though, KDE releases have all been pretty ok. 2.0 could have
    been a bit more stable, but it is questionable whether delaying it would have
    helped much.

    Instead of going in panic-mode it is usually more constructive to check for
    remaining problems and either fix them yourself, or report them to one of the
    lists. Based on such reports a release coordinator will be able to get an
    impression of the overall quality and make an informed decission whether to
    release or to postpone.

    Unlike popular believe there is no shame in delaying a release till it has
    reached a quality that is desirable. It is up to the release coordinator to
    decide when that point has been reached. The sole purpose of release
    schedules is to coordinate develoment _WITHIN_ KDE itself. Distro's like
    Conectiva, Mandrake or SuSE may find it inconvenient that a KDE release
    happens later than originally planned but quite frankly that's their problem,
    not KDE's.
    (And in my case that's partly my problem because I happen to work for SuSE,
    but I disgress).

    I would also like to make use of the opportunity to thank Dirk for his hard
    work on this release. Thank you Dirk!

    And now let's have some fun finding those last remaining bugs....

    Cheers,
    Waldo
    --
    Advanced technology only happens when people take a basic idea and add to it.
    -- Bob Bemer

  • Bah! (Score:5, Interesting)

    by Anonymous Coward on Saturday March 09, 2002 @10:10PM (#3136623)
    This guy has been moaning about KDE since ages. He may contribute to KDE a bit, but by God he pisses off nearly every developer.

    lists.kde.org and dot.kde.org are where he trolls most.

    He has:

    Criticized *many* KDE developers good work, even though they are working for free in their spare time.

    Would rather see Microsoft go off scott free and end up killing KDE than have Microsoft be punished for being a monopoly.

    Has sabotaged KDE CVS because people didn't agree with him.

    He now wants to lead KDE.

    The guy has an agenda to cripple KDE anyway he can, by sowing discord and criticizing everyone. He shows no respect for peoples work and never apologizes even when he is completely wrong. Its a miracle KDE has put up with him so far.

    • Re:Bah! (Score:1, Informative)

      by Anonymous Coward
      .. and not only that.. but his app (KIT), makes kinkatta not be in KDE.

      for everyone who has used both of them, kinkatta is a SHITLOAD better.
    • Neil contributes more than a bit. Half the Noatun modules are his, a new game of his will make 3.1, he works with me on a project to improve video support in arts. More importantly, he dares to express his opinion and is very serious about keeping KDE completely free and open in development).

      And how has Neil sabotaged KDE CVS in any way?
      • I could never get Noatun to work, so I ain't that impressed.

        I think the underlying issue seems to be how "dictatorial" the open source process should be. My view is that for a highly integrated project like a desktop environment, where other people will rely on the libraries you code, it should be very dictatorial. KDE has prospered because it has had tough rules on code quality. Gnome - I have heard - is slacker and has had problems. (No disrespect tho, I am looking forward to 2.0) The same with Linux itself, Linus has admitted he is a bastard, and his tough, blunt management style has worked well. Open source gives you lots of free code, but it is important to select the _good_ free code. That in turn gives the coders more self-respect as part of a disciplined project.

        Just my uninformed 2c.

    • *raise the scimitar*

      * Yup, plenty of people do get annoyed with me. But is that in itself a bad thing? I'm reminded of the Peter Principle, in the part where it's shown how bad processes get reinforced when those who make waves are tossed out.

      * Yup, I criticize things. I also beg for my work to get tested and criticized. What I *don't* like are vague criticisms with no hint about what to *do* about them. My criticisms on the matter of KDE 3.0 have not been vague, and certainly have included suggestions on what to *do* about them, even including my volunteering to handle it myself the next time around.

      * Microsoft is irrelevant to KDE.

      * Sabotage? I botched one import, which briefly cluttered kdemultimedia with an extra dir before Dirk fixed it. Yes, the same Dirk. Yes, I apologized for the error. Sabotage usually implies that something breaks. I broke nothing.

      * Lead KDE? Well, yes I said leader. When it's used in this manner, maybe leader was the wrong word. You need the context - when the RC proposes a schedule, anyone and everyone comments on how it can be improved, and it's the developers collectively who make the decisions. The RC just enforces the decisions that everyone else made.

      * Agenda to cripple KDE? I'm sorry, I just can't answer that one. I don't know where to begin.

  • I have to say it... (Score:2, Insightful)

    by xonker ( 29382 )
    Three developers out of all of KDE's developers criticize the KDE dev process and it's newsworthy?

    Now it's guaranteed to be picked up by other "news" sources and generate a bunch of useless controversy. Sometimes when I see what Newsforge (gag) puts out, I wonder if they have a secret agenda to discredit Open Source development. A quote from someone (Tina) at Newsforge about their policy of posting any and everything:

    we assume that our readers are smart enough to separate the sheep from the goats [tuxreports.com]

    I've gotten this response from them before, and they don't seem to understand the difference between filtering crap and saying that they shouldn't post critical or negative commentary. Yes, by all means, run a negative story if it's important, but don't run crap -- positive or negative. Posting a diatribe by three KDE developers -- folks, that's crap.

    In other words "don't look to us for news, because we don't do perform any kind of quality checking or the typical gatekeeper function you'd expect from real journalists. Someone submits it, we'll post it somewhere."

    I'm not saying that it's wrong to report on genuine conflicts or negative stories when they're important, but this really doesn't qualify IMHO. When a site purports to be a news site, there should be some quality checking and filtering going on. Stories that are comprised of nothing more than a rant or stories that are obviously biased or outright false should not be picked up -- or at the very least have a disclaimer attached.

    Microsoft must love the fact that Free and Open Source software development discussion takes place in the open where everyone can examine and dissect every personality conflict and internal bitchfest. It makes the Linux and Open Source community look like a bunch of fractious losers while no matter what Microsoft PR's department spews everyone marches in lock-step to it.

  • have a document like the Constitution where we can abolish the management if it tries to take out our fundamental rights to get non-buggy software? Cause, I'm available as an asassin if necessary. ;)
  • KDE 3.0 is currently in beta. This means that if it doesn't compile or is missing things then so be it. Beta software by definition isn't ready for general release. (I know redundant but you know...eventually the point may sink in)

    The 2.4 series kernel and recent Patch penguin broohaha could be criticised for the same. In fact the 2.4 kernel has been called the "Kernel of pain" because of the problems. What this has done is force the development community to work harder, and even Linus is trying new things (automated patch integration into the tree). I believe that constructive criticiscm is always helpful, and a good sanity check BEFORE KDE 3.0 is released does not in any way take away from KDE. Perhaps the can avoid the pitfalls of the 2.4 series Kernel.

    • Neil here,

      bzzzzzzzzzzzzzt. KDE 3.0 is *not* in beta. It has passed the beta stage, at least officially. It is a release candidate, which is something that is supposed to be release worthy.

      Don't believe me? Before the Nuernberg meeting, the schedule had KDE 3.0 final coming out just a few days after RC 1. That's how bad a failure RC 1 was.
  • That is as the API goes, spend a few hundred hours making your apps work on kde2, a few months later have to spend a few hundred hours making your apps work on kde3, a few months later....... This is bad for kde and bad for linux.
  • in my experiences in release management over the past 15 years, i would say unfortunately, this story is the norm in Silicon Valley software companies (during which time i worked at two of the largest unix system houses in the industry,
    and other smaller campsites)

    over the last 10 years, i've observed that fewer and fewer teams are able to meet schedules. one of the biggest problems faced by todays teams is that the time to market have been reduced significantly, yet the infrastructure needed to support these teams is still dependent on tools and practices from 20 years ago...

    another key problem that large product teams face, is that the software is not architected to support the kind of rapid change that is required to compete in todays markets. this results because of the fact that the software industry has architects of technology, but are lacking in architects of product.

    when was the last time during a software design review, anyone has seen how the product was going to be supported, and what impact that has on the architecture...

    Regards,
    Kramer

  • Neil here (Score:3, Interesting)

    by JamesKPolk ( 13313 ) on Sunday March 10, 2002 @01:49AM (#3137044) Homepage
    Lots of interesting comments here...

    I'm glad people have noticed that this isn't a criticism of KDE, or the direction KDE's taking or anything. It's just a criticism of the breakdown of the policies that were followed for the 2.x releases.

    Some have questioned whether this is even newsworthy. I figure it is in the same way an internal Microsoft memo would be considered newsworthy: It's the fact that it's one snapshot of things people don't normally follow that makes it interesting.

    And of course, conflicts make headlines.

    I encourage all users and fans of KDE to make a good backup and test KDE 3 RC 2, following the tips on http://www.kde.org/kde2-and-kde3.html and http://www.kde.org/install-source.html to help make KDE 3 the best release it can be, whether a beta 3 is released or not.
  • I don't think anyone mentioned this, but 3.0 is called that for a reason. This is why we have 3.1 and 3.2, etc... afterwards. So right now we are at RC for 3.0. Big whoopdeedoo. I'm just a casual user right now. Actually I'm kind of a power user, I just have no need for bells and whistles and I'm not a developer. I plan on sticking with 2.2.2 for a while since it is so stable, and wait until 3.1 comes out anyways (unless I hear the 3.0 is stable when it comes out). dot.zero releases are supposed to have problems with them. The guys who posted the original article to the mailing seem to be confusing their argument between things that are wrong with KDE itself, and things that are wrong with management of KDE. I don't think getting a new leader will change much. Open source projects like KDE rely on hundreds of developers and it is a team effort. Under a different leader the same problems would still have occurrred.

  • Read the fine print (Score:2, Informative)

    by CanadaDave ( 544515 )
    I think if those three guys have of read this and understood it, maybe they wouldn't be ranting and raving about the problems with 3.0

    Taken from 3.0 release schedule [kde.org]:

    What's the deal with that feature-plan?

    In the past, there were a lot of complaints about a rather long "freeze period", so this is an attempt to address this issue. Basically the idea is that you add an entry about what feature you want to finish in the 3.0 timeframe and mark it as done when you completed it. This helps me to get an overview about the outstanding changes and in return allows you to work on the missing parts even in the "freeze period". The feature-plan is open for commits till November, 2nd. Later additions require reviewal first. I will try to respect outstanding entries in the release schedule. Please understand that although this gives you greater freedom over CVS, it also requires more discipline in making sure that your additions don't have unwanted side effects.

    I think what's said above can explain most of the bugs and other problems that still exist in 3.0 Beta/RC

  • One very important point (Score:4, Insightful)

    by forgoil ( 104808 ) on Sunday March 10, 2002 @04:55AM (#3137273) Homepage
    There is one thing in their mail that really struck a cord within my development genes. Because there are a host of important proceedures, and these apply just fine to any kind of development, regardless of it's Microsoft or KDE. This is also one of the reasons why I was so incredibly impressed with KDE.

    But as the letter proves, they have now failed in two incredibly important ways. First of all they don't manage package upgrades well, which is a mortal sin in my book. But further more, they absolutely number one sin is to commit anything to the reposatory that doesn't build. This is a very fundamental way of working, you will waste a large number of hours if the central build doesn't build. All of a sudden everybody has to fix this somehow, and just for the sake of argument say that 50 developers got a broken build. They find the problem and fix it (I know that some might wait for a patch), there you have 50 potential other bugs that could have been fixed instead. Make sure everything does build before you commit. Use branches and make it safe.

    And even worse, 100s of users that wants to live on the bleeding edge and help find bugs are all of a sudden faced with build problems (which can be avoided, there are commercial projects that excel at this. There is nothing that says that free projects can't as well). Wasted hours and sour faces for everybody.

    Give some (light) punishment to the last person who broke the build, this works, and motivates. One just can't make a true punishment, because that is quite frankly stupid. Something like having to answer user mail until the next one comes along and breaks it or something along those lines.
    • your point is very well taken. in my experiences in release management, i've found very few development teams that are able to achieve successful daily builds... there is a blend of symptoms: know-how and team-culture are the
      two areas that need to be addressed.

  • I noticed this complaint reported on Sourceforge yesterday. All I can say is, based on my experimentation with RC2 and relatively recent CVS's, things don't seem to be anywhere near as bad as the complaint implies.

    Realistically, what I found was only ONE serious bug that keeps me from using the KDE3 CVS release as my current desktop - and the reply to the complaint mentioned it:

    Except for khtml problems I would say KDE is in a pretty good shape right now

    The big problem right now, from what *I* have noticed (there may be others, but I haven't stumbled on them) is the broken focus. I couldn't write this post in KDE3, for example, because while in the textarea, the "focus" is actually still on the links in the page. Pressing the Enter key while typing here in KDE3 would cause the browser to jump to the currently focussed link (the first one on the page) instead of putting an "enter" into the textarea...

    While the fact that this huge focus problem has been in khtml for so long and (as far as I can tell from what I get out of "cvs update kde" from anoncvs.kde.org) isn't being addressed at all disturbs me (bugs.kde.org now even has a bug entry set up to track all of the 'duplicates' that are all permutations of "keyboard focus in khtml is broken"!), the fact is that other than this ONE bug (which may conceivably only still be there because of the feature freeze preventing a reworked set of khtml focus code from being committed to fix it), KDE3 was looking like it was actually in quite good shape...

Slashdot Top Deals

news: gotcha

Close