LWN: Comments on "KDevelop 4.0 released" https://lwn.net/Articles/385888/ This is a special feed containing comments posted to the individual LWN article titled "KDevelop 4.0 released". en-us Mon, 08 Sep 2025 07:18:16 +0000 Mon, 08 Sep 2025 07:18:16 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net KDevelop 4.0 released https://lwn.net/Articles/386201/ https://lwn.net/Articles/386201/ jmspeex <div class="FormattedComment"> Well, among the things they dropped was automake integration, which I happened to be using. Also, kdevelop 3 was abandoned for so long before Kdevelop 4 was released (assuming it's usable now) that I ended up switching to Eclipse. I'm not looking back. Also, it's not like they dropped one or two features. This is a completely different piece of software. They might as well have used a different name. <br> </div> Wed, 05 May 2010 03:46:26 +0000 KDevelop 4.0 released https://lwn.net/Articles/386044/ https://lwn.net/Articles/386044/ sebas <div class="FormattedComment"> GNOME 2.0 at least came with *significantly* less features than the last 1.x version. That's fine and dandy, as long as people manage to read release announcement instead of just projecting their view of how things should be onto everything else (and in the process conveniently forgetting about verifying their input data). A .0 project also often is the result of refocusing the goals of a project, in that sense, you're bound to remove features that don't contribute to your goals, that's part of the process.<br> <p> Wether KDevelop4 is finished or not is in the eyes of the beholder:<br> - For C++ developers, it's definitely a stable release<br> - For PHP developers, the same holds true<br> - Who expects an IDE to support any language equally well?<br> - sometimes, not always, less is more, it's even very UNIXy<br> <p> If you want to create a complex piece of software, such as an IDE, and expect yourself and your team to have everything in place by the time you release, you've most likely lost your momentum way before you come even close to what you envision as a release. Perfect is the enemy of good.<br> <p> It's absolutely not uncommon for any kind of software to focus on a limited set of features, get that right and stable and then think about extending the feature set.<br> <p> </div> Tue, 04 May 2010 11:27:41 +0000 Another complete rewrite? https://lwn.net/Articles/386043/ https://lwn.net/Articles/386043/ sebas <div class="FormattedComment"> On one hand, Qt4 is quite different from Qt3, on the other hand, KDevelop3's architecture had reached the end of its lifetime, and bigger new features (such as the pretty awesome semantic highlighting that it has now, but also UI changes) would've been hard to implement.<br> <p> Some developers have therefore decided to have The Big Rewrite coincide with the Qt4 Porting (KOffice and Amarok did the same, btw), others have decided to first to a pretty plain KDE4/Qt4 Port, and then start reworking the architecture. Kontact, Digikam and K3b would be example for "relatively straight-forward" ports to KDE4 architecture, with Kontact now being ported ot use Akonadi, ETA is KDE SC 4.5 this summer.<br> <p> Both approaches have up- and downsides of course, what's best is usually at the discretion of the individual developers and needs to be decided on a case-by-case basis.<br> </div> Tue, 04 May 2010 11:20:01 +0000 KDevelop 4.0 released https://lwn.net/Articles/385987/ https://lwn.net/Articles/385987/ cowsandmilk <div class="FormattedComment"> Keeping every feature around is often counterproductive.<br> <p> It's a development environment. If developers making the software never use a certain feature, it might indicate it isn't needed. Or if they implement a feature they don't use, it likely won't be done optimally to suit the needs of the users. Instead, it seems they focused on making the features people did use complete and half-assed, which I can applaud.<br> <p> Additionally, among those features they did not implement were Perforce and ClearCase support, instead adding git, mercurial, and bazaar. Focusing on the DVCS regime is a clear win in the eyes of most open source hackers. And removing support for features that would otherwise just result in cluttered UI is great.<br> <p> I think they should be applauded for thinking about how to develop a better product, focusing on what they would want out of a product used for development, and making a product that gets those things done.<br> </div> Mon, 03 May 2010 22:31:35 +0000 KDevelop 4.0 released https://lwn.net/Articles/385954/ https://lwn.net/Articles/385954/ Kit <div class="FormattedComment"> I don't really see that as irony so much since it's an IDE meant first and foremost for C++ development.<br> </div> Mon, 03 May 2010 19:48:21 +0000 KDevelop 4.0 released https://lwn.net/Articles/385941/ https://lwn.net/Articles/385941/ kragil <div class="FormattedComment"> So they miss a few languages. Big deal, those are plugins now. Write them if really care.<br> <p> But really, it now has a ton of new really advanced features and I really don't see the need to start the stupid old boring KDE 4.0 discussion again. They worked for years and years on this and had betas and RC for years. That certainly is good enough by any standard. <br> Nobody cares about the messed up 4.0 anymore. It is history, deal with it and move on with your life.<br> <p> EOD for me, but before I forget: <br> Congratulations for this great release to all the developers.<br> </div> Mon, 03 May 2010 19:01:01 +0000 KDevelop 4.0 released https://lwn.net/Articles/385924/ https://lwn.net/Articles/385924/ lacostej <div class="FormattedComment"> The main problem is that a x.0 release is (in my mind) supposed to be ready for public consumption and has a (de facto) history of not losing major features. Apart from KDE 4.0 (which caused some to be very unhappy about), I know very little FOSS projects that released incomplete (featurewise) stable .0 releases. Gnome, open office, X, most Linux distributions, etc come with more not less features at .0 releases; missing features are typically because they are deprecated or conflicting with a new one, not because "we didn't get the time to adapt all our core features to our new framework".<br> <p> I understand the desire to get something out of the door and tested by many, but if it's not finished (from an end user perspective), expect a communication problem.<br> <p> Alpha refers to quality, but also to unfrozen status, that's why I used it. It's more appropriate than beta in that regard. What are the alternatives ? (n-1).9[9] ? (e.g. 3.99 ?)<br> </div> Mon, 03 May 2010 17:25:23 +0000 KDevelop 4.0 released https://lwn.net/Articles/385918/ https://lwn.net/Articles/385918/ MisterIO <div class="FormattedComment"> The irony is that, by default, it doesn't even support C, only C++(at least if it's not very different from v4.0RC3)!<br> </div> Mon, 03 May 2010 16:31:43 +0000 KDevelop 4.0 released https://lwn.net/Articles/385917/ https://lwn.net/Articles/385917/ AndreE <div class="FormattedComment"> Alpha/Beta/RC generally refer to stability and quality. <br> <p> Being a stable release but lacking some features seems perfect for a x.0 release<br> </div> Mon, 03 May 2010 16:27:29 +0000 Another complete rewrite? https://lwn.net/Articles/385915/ https://lwn.net/Articles/385915/ AndreE <div class="FormattedComment"> yes? (from what has been said)<br> </div> Mon, 03 May 2010 16:22:24 +0000 KDevelop 4.0 released https://lwn.net/Articles/385909/ https://lwn.net/Articles/385909/ lacostej <div class="FormattedComment"> Why not call it 4.0.alpha then ? If you remove so many features ? Or are other languages optional ?<br> </div> Mon, 03 May 2010 16:03:54 +0000 Another complete rewrite? https://lwn.net/Articles/385900/ https://lwn.net/Articles/385900/ proski <div class="FormattedComment"> Why is KDE so obsessed with complete rewrites? Is Qt4 so different that the original code cannot be ported?<br> </div> Mon, 03 May 2010 15:48:40 +0000