A slick KDE Traffic #53 is out. This one features a single long-standing topic, the release cycle of KDE 3.2. Read all about it at the usual location.
Dot Categories:
A slick KDE Traffic #53 is out. This one features a single long-standing topic, the release cycle of KDE 3.2. Read all about it at the usual location.
Comments
Thanks for the pointer! This is just too cool. :)
>qt-copy... installed at the same time...
How does that work? How to set it up?
Derek
The thing is that kde project only releases source-- KDE is supported in so many platforms that it's impossible to make packages for every single one. So, in every stable KDE release, third party packagers are given the source early (typically a week ahead or so), to package it. It's a bit of a burden for these packagers to have to package development releases.
In the gnome side, it works because you have companies like Ximian who employ people to package stuff (such as the dozen different linux flavors supported by redcarpet..) Companies that sponser KDE development, such as SuSE, typically are only concerened with their distro only (and rightly so, why would SuSE start releasing redhat or Solaris packages? :))
For those who really need to live on the edge, grab yourself a debian unstable then read: KDE CVS HEAD Debian packages
This would mean: making sure it compiles, branching, freezing and pre-packaging, checking it's even coeherent and starting again. Perhaps it could took just some hours or even some minutes a month, but these are hours or minutes much better spent coding and bugfixing!!
Maybe if anyone offered just to do this, but again who would benefit?
On users' side, more users who don't understand programming enough to configure/make coud install some crashing unstable realeas (what a deal!)
On developers' side, coders would drown under a lot of quite untrackable bugreports due to the proliferating of different versions, subversions and snapshots.
Just my 2eurocents.
As a Solaris administrator I'd prefer less frequent releases. KDE isn't as ready for Solaris as it is for Linux which invariably means I have to spend more time compiling the latest release.
Furthermore, there is no good upgrade path. New version means a new directory tree. That means I have to deal with hardcoded file paths in some of the preference files. And then I have to figure out all the changes I made and reimplement those, such as what I moved out of the global autostart, which screensavers I disabled, etc.
As far as I am concerened the word "upgrade" is a misnomer. "Migrate" would be more correct and it would be really nice to see a migration tool created to handle this.
Regarding the recent threads of discussion regarding the release schedule of KDE 3.2
For the most part, I am just an observer. I have tried Linux and KDE and do thoroughly believe that it will be the future and certainly is deserving of use within many workplaces.
My comments about the release schedule are: KDE may be getting too big to release all at once.
Let me explain. There are many parts of KDE that should be bundled together and released together but there are other parts that could be separate. I remember when Windows was a pretty basic operating system (environment) but with each new release, it has more and more features that were copied or acquired from other developers. With regard to KDE, I think that perhaps for version 4, there should be a separation of certain parts of it, each with their own release schedule. For example, games, KDEPIM, KMail/Kontact and perhaps some other "packages" could be split off and released separately.
In splitting off some of these items, it may enable the KDE core to move more quickly from release to release. Furthermore, it will enable the split off packages to achieve autonomy from KDE and release on their own and not have to wait for the rest of KDE to reach release level. A spin-off from independent releases is that with different packages being released at different times, it will give the appearance that KDE is just-a-hopping with a new release of something or other every couple of weeks. Of course, when the next major release is being prepared, presumably because of major underlying changes, each project may be timed to have a release at the same time.
Please excuse any ignorance that I may have revealed about my knowledge of KDE: I am just putting in my 2cents.
Jules
Thoughts on what's missing, needed and new directions on KDE
I have been reading a discussion on which direction the KDE or new features the KDE should move in on. I choose KDE over the other window managers for one simple reason; better integration. I can see two directions for KDE to move on in; better office integration, hence kontact being the major application on that. KDE lacks something like Evolution; KMail is a very nice tool but has to be integrated with address book, organizer. That will complete the package for a business or a home user enabling it to reach all those commonly accessed tools from one place.
Although KDE is still mature, KDE control panel is still too much convoluted. Certain things have to be compacted into one. There is an option for fonts, colors, icons, window decoration, style, themes. A theme should be composed of icons, colors, window decoration + style. Panels is an unnecessary feature, I doubt many people use it hence should be hidden. I think a couple of months should be devoted to usability study and see what is most commonly used and pushed to the front so that a novice user can gain the most by spending the least time with the control panel.
Also, integration in konqueror should be in the forefront. Especially the kmplayer should be enhanced as much as possible giving the user a polished feel rather than a clunky working plugin. Right now, I am happy with kmplayer a lot but more can be done.
I would like to thank many developers for putting so much of their time to bring such a package together. Many thanks...
Note that kmplayer was original just a demo app for a khtml plugin with javascript bindings, as a kde alternative to nspluginviewer.
All that was added later was for enabling mm devices I have at home. And the broadcasting was added so that I can watch tv at diner :-)
What you call clunky look, is of course a personal taste. I just don't like big gui elements.
But, quite frankly, I don't believe kmplayer will live after one or two releases. Arts with xine can play as well as mplayer (and maybe even better for real streams) and it's fully intergrated into kde. As soon as apps get released capable of playing from all the sources kmplayer does (qtvision comes into my mind, if it weren't in beta forever, and/or noatun/kaboodle), I stop working on kmplayer.
Panels is definitely very useful. I have a multimonitor setup and being able to set some better settings for the panels under that situation is very nice. I don't like the idea of hiding options since in the end it just makes systems harder to use.
Overall kde is still the only environment I have used that actually works well with xinerama,
But if you don't like the panel, aren't there options to hide it? Nothing is visible on my desktop except for my wallpaper. If you think the panel is too cluttered, you can remove and add certain items as you wish, and even configure the behaviour or properties of each item. I did just that. Again, this is the power KDE allows its users, the choice to metamorphosize your desktop into what you want. With regards to KMAIL, I think Kmail should remain what it is,an email client. KDE comes with an address book and an organiser, why would anyone want to further clutter a program when it can run faster and better while it is lean. Not to mention it's more manageable.
Take a look at Microsoft Outlook. Half of the problem encountered by that email client doesn't stem from the email client itself, it stems from redundant and unnecessary addons such as organisers, schedulers, alarms, address books etc. More integration means a slower app, more bugs, and the possbility of a latent security hole. Not something we want in an email client.
Like you, I do agree and hope kmplayer is better integrated into Konqueror and becomes the default media/video player. Aktion and kaboodle are just pathetic. They are an embarrasment to KDE's caliber. They should thrown in the trash can and lost there forever! I think we should spend much more time integrating (k)mplayer, which is much better than Xine, into KDE/Konqueror. If properly implemented, a part of KDE's multimedia woes are over.
Regards,
Mystilleef
There is a mailing-list dedicated to KDE usability, why don't you look at the archives ? There was some discussion about themes and I have started to develop a new control center module... Just wait a few weeks and it will be ready.
> new control center module... Just wait a few weeks and it will be ready.
aaaahhh! tasty! :-P Willing to see kcontrol broken again ;-)
???
Sorry if I offended you or my bad english was uncomprehensive. :-S
What I meant is that whenever new features are added to kcontrol something breaks (for example, it's not long that I found kcontrol crashing when I used big icons instead of the small icon tree).
Anyway, I like to see kcontrol broken sometimes. That means somebody is improving it :-)
The kcontrol shell shouldn't break when some kcmodules are getting tweaked or exchanged, both are pretty independent from each other. Most likely you are talking about compile or packaging problems.
>The kcontrol shell shouldn't break when some kcmodules are getting tweaked or >exchanged,
Errr... when I said that "it breaks" I didn't mean it broke the compilation or linkage. I meant it breaks the program, i.e., crashes. Obviously, if a module is not dealing properly with memory and segfaults, the whole kcontrol will crash. Well, at least I never saw a module crashing itself and leaving kcontrol still working. Maybe that all those bugs were due to kcontrol then.
>Most likely you are talking about compile or packaging problems.
No, I'm talking about a recent CVS Head version. That's an upgrade of qt-copy +kdelibs+kdebase on the 24/May/2003 around 1:15 am. ldd reports no reallocation problems, and all apps run fine, but kcontrol has crashes.
I open kcontrol, Icon View mode. No problem. Click System Administration, no problem. Appearance->Theme Manager crashes. The same shortcut *doesn't* crash in tree-view.
So it seems that there's actually a connection between kcontrol and its modules, since it crashes only with certain modules, and icon view mode in kcontrol.
How about something as simple as: feature freeze the last 5 days of a month, coinciding with a big bug-fixing party (real or virtual), and release whatever turns out on the 1st of the next month as a preview release.
This would allow developers to keep tacking on features all the way until the official feature freeze, but have a dedicated time in which to say "okay, let's clean up what we have." The result doesn't have to be perfect or totally devoid of grave bugs to be an improvement.
Apologize if this has already been mentioned but I didn't see it.
Nice idea, but I don't think it will work, simply because most contributors to KDE do the work they do when they feel like doing it, not when some schedule tells them to do it.
I think this happens naturally. A bunch of new features will go in, then everyone will scramble to make it work. Then a new wave, etc.
It isn't documented, or planned, but happens.
derek
Bug fixes vs. Features, Features vs. Bug fixes. And then a false compromise. :-|
I do not see this as an either or proposition.
I posted this idea before and nobody liked it but:
The bug fixes and new features should not both be done on the same branch!
Bug fixes should be done on a feature frozen branch (i.e. the current release) and new features should be developed in their own space separate from the bug fixes.
As I said before, new features should accommodate bug fixes no the other way around as it is now done.
With this paradigm shift:
1. Some time would be saved because it would not be necessary to: "back port" bug fixes.
2. Some additional time would be spent because bug fixes would probably break some new features.
--
JRT
James,
If I remember correctly, you were the author of the post that I referred to up at top, that I've been avoiding commenting on because it was frankly not really to anyone's benefit for me to do that. And I'll avoid commenting on that specific post at this point, but I think I would like to make some general observations that may help you in your interactions with the KDE people.
Treat them like people. If they don't like what you posted, give the reasons that they give serious consideration, and respond to the points as if they've made a valid argument, *even if they haven't*. Take their point of view seriously, even if you think it's the stupidest thing you've ever heard. Don't look down on them because they may have less of a formal education of less "experience" or what not. Personally, I don't have a college degree, but I have quite a few years of programming and administration experience under my belt (close to nine at this point), but you'll notice this is really the first time I've mentioned it other than by way of introducing myself. A lot of these guys are young and just starting out in programming - don't begrudge them the satisfaction of working on code that they believe in.
I would like to preface this with the second point with the caveat that I have been disabled myself and know what it is to want to contribute something - anything - but not being able to. It isn't a pleasant experience. But the general culture of open source software tends to be "put up or shut up". You either put in what scratches an itch,ask nicely for someone else to do it (chances are they will if they think it's a good idea), or suffer in silence. The people who are writing the code - for free, I might add, and in some cases at a great personal sacrifice to their lives, such as they may be - aren't interested in someone who hasn't proven their mettle coming in and telling them how to do it. Even if they may be more experienced. This is a meritocracy.
I know that you may think that you have answers and that no one is listening to you. Step back for a moment and think about *why*. Is it your approach? Is it just that you haven't proven yourself?
I applaud your wanting to contribute to the KDE project, in whatever way that you can. I do think that your approach leaves something to be desired. Have a *conversation* with these people - and treat them as equals - and I bet you get much farther than you would otherwise. I have a philosophy in life, myself - even a PhD, even though it may help you to get a better, higher paying job than you would otherwise, still only means that you have proven your ability to bullshit a review board. Even someone with multple doctorates has to use the bathroom twice a day.
Hope this helps. And if you're not the guy I'm thinking of, sorry.
--Russell
I thank you for your thoughtful (and somewhat long) comments.
Yes, I plead guilty of making somewhat sarcastic and terse remarks to people that anonymously posted what I thought were inane comments.
Yes, some of what I said were "inside comments", intended as humor, but probably not take that way by those that didn't get the joke.
> [They] aren't interested in someone who hasn't proven their mettle coming in
> and telling them how to do it. Even if they may be more experienced. This is
> a meritocracy.
And, I had this strange idea that my ideas might be considered on their merit. :-)
I remain frustrated that there was no actual discussion of any of the actual idea. Or, did I miss something?
The only response to my original posting on kde-devel was that I was told -- to coin a phrase -- that it was the stupidest thing they'd ever heard.
I have put the idea out in what I think should be a good forum. We have had the comments of several users supporting my basic premise that fixing bugs is more important to users than new features. So, it would be nice to discuss the idea rather than discussing me.
What do you think about it?
--
JRT
I'm well known for the amount of verbiage that I can spew. I'm one of those (and probably thankfully) rare people for which each word has a meaning, and different words that are close in meaning can still have subtle shades of meaning that most people don't pick up on or care about. So in an effort to give the most precise meaning possible to my words, I end up using a lot of them. *shrug*. I've grown used to it. To me it is somewhat similar to a fine musical performance or elegant computer code - just in language.
I think that we have slightly different meanings for the word "meritocracy". Perhaps I used the wrong word in this case. The way I meant it was "by he merit of what you produce" rather then "by the merit of your ideas". Ideas are wonderful but without implementation, they are just ideas, and it's impossible to force volunteers to implement ideas, no matter how well thought our they may be. There are several programs that I would like to see, myself, but I don't even bother asking for them, as I realize that if I don't code them myself they will not get coded. That's regardless of how useful they would be.
Anyway, what do I think of your proposal. I'm not sure what I think of it. I am a developer on one level, yes, but I haven't the experience with large projects that most of the other KDE developers to - my experience tends to be as a "lone ranger" type coder. So I really can't give you any constructive feedback. As you can probably tell, I feel my strengths lie somewhat more in the "people" area, and fortunately, I've found a way to contribute using that without seriously rankling people in the process :-) I know what I want in KDE, I can help put it there if I feel strongly enough about it, but I'm not particularly interested in all of the guts of project management.
--Russell
> I think that we have slightly different meanings for the word
> "meritocracy". Perhaps I used the wrong word in this case.
I was just making a facetious remark. Your meaning is basically correct as far as M-W is concerned:
1 : a system in which the talented are chosen and moved ahead on the basis of their achievement
2 : leadership selected on the basis of intellectual criteria
The problem with this is that this definition doesn't rule out the achievement of a college education.
So, the meritocratic system you describe would be a special subset of a meritocracy for which there is probably not a specific word.
My point here is that if someone says that my college education doesn't matter because this is a 'meritocracy' they are saying something that may make sense to them, but strictly speaking they are actually contradicting themselves.
--
JRT
You are probably correct.
I propose we coin the term "codocracy" :-)
--Russell
In my office environment, we work daily with most of the Outlook features. A switch to KDE-Linux is only possible if the promised kroupware integration and Exchange comptitbility is there.
In short: forget about success on our deskop until 3.2 is there and stable. It seems to me urgent and important. A feature freeze seems to me necessary. If there are problems with kroupware, as this is a priority for success, they should be sorted out before new gadgets are added.
Please thanks for all the wonderfull improvements and staility added since 2.0.
Geert