Phonon - Qt or KDE?
tdfischer at fedoraproject.org
Fri Nov 11 17:59:17 GMT 2011
Since we can't just sit on this discussion, here's a list of pros and cons
I've made in the case of moving from KDE to Qt
* Closer to Qt developers
* Helps everyone forget about that QtMultimediaKit blunder
* Can pull in more Qt users
* Can help find any big bumps or rough spots for other projects who would
like to perform the same leap
* Not everyone hates it
* Need to update all our URLs. Again.
* Possibly involves getting a new mailing list
* Some KDE devs have an irrational fear of working with Nokia in this
manner, much like people don't like using pgst because gstreamer isn't a
KDE project and we hates the gee oh yes we does, precious.
* Not everyone likes it
* Really, the shed should be azure with green trim.
Finally, here's some arguments I've heard that I find a bit specious:
> We'd be moving from one community to another
Not really, since the whole goal of KDE Frameworks and open governance is
to bring everyone together by merging large swaths of KDE code into Qt
proper. There isn't any movement of people. I'll still use and contribute
to KDE for many years to come. KDE folks can still use Phonon, and Qt
folks can still use it as well. There isn't a movement at all, merely a
relocation of where people point git and bug reports at.
> Nokia hasn't treated us properly in the past
Qt isn't Nokia. That statement makes an implication that we are somehow
giving up Phonon to Nokia/Qt which is blatantly false.
> We'd have to port everything to use qmake instead of cmake
All build systems suck. CMake just sucks less than others in some places.
If qmake supported pkgconfig natively and supported the generation of .pc
files, I'd probably actually say thats a huge incentive to use it. I'll
save that discussion for another day though.
> Lo' all!
> Some time ago I got asked whether we'd want to move Phonon to the
> qt-project infrastructure to reflect that it is not only for KDE
> software but the wider Qt ecosystem.
> That would mostly include:
> * gerrit for code review
> * jira for bugs
> * whatever is going to be used instead of gitorious for hosting (FWIW:
> moving back to gitorious is a no go ;))
> Disadvantages that come to mind:
> * KDE devs will need to get a new account to contribute to Phonon
> * bugs in KDE software that are really in Phonon cannot easily be
> moved within bugzilla but need to be reported on jira
> * the recent merge of #phonon into #kde-multimedia would seem somewhat
> weird? (thought I suppose we could get #qt-phonon ;))
> Advantages that come to mind:
> * easier to market Phonon as *the one and only reasonable* Qt multimedia
> Additional note on closeness to Qt:
> * as Qt 5 will not contain a Phonon module anymore, but instead
> developers are supposed to get it from us directly I am strongly
> considering to move Phonon to qmake and reduce the amount of
> not-necessary deps (automoc4 for example, which is cut by qmake itself
> already) to make it *a lot* easier for people to get the latest
> version and develop against that
> * on a highly related matter I am not entirely opposed to the idea of
> using qdoc3 instead of doxygen (which gives the advantage of writing
> actually sane QML documentation and extremely magical theming for api
> dox + our doxygen stuff on api.kde.org has been broken for the last 3
> years and no one really knows why it seems)
> So, what do you think? Move to qt-project or stay on KDE infrastructure?
> kde-multimedia mailing list
> kde-multimedia at kde.org
Trever Fischer (tdfischer)
Fedora Ambassador, KDE Hacker
GPG: C40F2998 hkp://wwwkeys.pgp.net
More information about the kde-multimedia