Linked by Thom Holwerda on Mon 21st Jan 2008 18:22 UTC
Ars reviews KDE 4.0.0: "KDE 4.0 was officially released last week after extensive development. The long-awaited 4.0 release ushers in a new era for the popular open-source desktop environment and adds many intriguing new features and technologies. Unfortunately, the release comes with almost as many new bugs as it does features, and there is much work to be done before it sparkles like the 3.5.x series." They were also at the KDE 4.0 release event.
Permalink for comment 297239
To read all comments associated with this story, please click here.
To read all comments associated with this story, please click here.
Member since:
2005-07-07
I am not arguing the release early, release often software development model - I am arguing the KDE team's implementation of this model. The release early, release often model does not mean: release a gold release as soon as possible - no, it means: release early in the development process. There is nothing in the RE/RO model that mandates that you should get a .0.0 release out the door as "early" as possible - it just mandates that the development process should be marked by a steady stream of snapshots, alphas, betas, and release candidates during the development process leading up to a final release (the 'often'), a stream that should start right at the beginning of the development process (the 'early').
But Thom, you are arguing against release early, release often. I'm in no way affiliated with KDE, but I did work (on my owns small bit of) a very popular and rather big Free Software project. Release early, release often does mean spitting out a new version for mass consumption, even if it has bugs. Good enough, is good enough.
Being open source doesn't excuse you for bugs, missing features, or feature regression. What you're missing in your comparison is the role of distributions in Linux vs Apple or Microsoft. With Linux, the 'Monolith Product' is the distro release--- Individual projects, if they want to survive AND flourish, must release early and often. (With the only exception being the most mature of utilities. Things like venerable, command line apps that do one very specific thing.)
Projects that don't release die. It can be a lingering death, but it's dead non the less. That's the real value add of having distros making their own stable releases-- Even if you disagree with their choices, you get an 'official call' on if 'good enough' is 'good'.