Proposal to plan for "Milestone Releases" on the way to KDE4

[Alexander Dymo]

On Wednesday 25 January 2006 23:11, Richard Dale wrote:
> At the moment there is very little work going into the next release of
> KDevelop 3.4 due in a few months, when it really should be seen as a
> flagship app and marketed up there with Plasma and so on. Why aren't we
> adding great support for Qt4 and KDE4 development into KDevelop 3.x, like
> Qt4 Qt Designer integration?

On Wednesday 25 January 2006 23:11, Adam Treat wrote:
> Because KDevelop suffers from a terrible problem.  None of us have made
> KDevelop our priority app.  Most of the KDevelop developers work on other
> things or have other things/priorities going on.  It is not a good
> situation. And I agree, KDevelop and our KDE development tools do not get
> nearly as much attention as they require.  These apps and tools really
> should be the best of the bunch, but they are suffering from neglect.

I think KDevelop and development tools are suffering because they was
not positioned correctly. Two years ago I was overenthusiastic about the
possibility for KDevelop to become the universal IDE for all possible
development tasks. Looks like I wasn't alone and eventually the whole
team put a large amount of effort accomplishing those tasks.
I personally put lots of effort into making KDevelop a development
platform top build any kind of plugin-based applications.
But it looks like we paid too much attention to multilingual, plugin-based 
architecture and forgot the main goal.

I'm not going to say now we did wrong all that time. Certainly KDevelop
has tons of useful features now and KDevelop Platform attracted Quanta
developers and made possible Kuanta to be born.

But in fact we paid too much attention to the architecture and platform 
and little attention to the "usability" (usability as in usable for
development tasks, not usability as good GUI).
I've been watching at core KDE developers for three years. They still use 
vi (emacs). Why? Looks like KDevelop doesn't have clear benefits as
a development tool for them. This is, of course my opinion and I'd really
like to hear from KDE developers what they think is wrong in KDevelop.

But I have my own experience that tells me that KDevelop is usually
not a best tool for my development tasks. It's a cool IDE, but not for me.
Wtf? I participated in its development, I implemented many things and
now I don't use it much. This question bothers me every day and I'm
going to get an answer from myself at the first place.

Heh, I know the part of the answer. I need a convenient tool to "just do the
job". KDevelop 3.x is not there yet. It doesn't fit my workflow and IMHO
it doesn't fit the workflow of many KDE developers. 

I think we should make the next KDevelop not process-oriented, tool-oriented
or framework-oriented but rather developer-oriented and workflow-oriented.
Then it will IMHO gain more attention from developers. We need to make them
feel that there's an IDE that works for them in the way they want.
Unfortunately now it feels like a developer is employed by the IDE.
And this feeling is true for almost all IDEs out there...

PS: please don't think I'm pessimistic ;) I just work as hard as I can to
formulate and lay the foundation of KDevelop4 and I'd like to hear from
any developer who cares about KDE development tools their
needs, typical workflows, use cases, etc. This will help us to create an
"developer-oriented IDE" of my dream :)