FESCo meeting summary for 2009-06-26

[Matthew Garrett]

On Sun, Jun 28, 2009 at 11:35:07PM +0200, Kevin Kofler wrote:
> Matthew Garrett wrote:
> > The reality is that KDE *is* a second class citizen in Fedora - it
> > doesn't get anywhere near the attention that Gnome does.
> 
> <SARCASM>Thanks</SARCASM> for insulting our (KDE SIG's) work yet again,
> that's <SARCASM>really appreciated</SARCASM>! :-/

I think the KDE SIG does a wonderful job given the resources available 
to them. But implying that they're able to produce the same quantity of 
useful work as the Fedora developers working on Gnome is either 
unrealistic or a direct insult to those developers. Given the same 
average level of competence (and assuming useful management), a larger 
team will produce more work in a given period of time.

> Where are the monthly bugfix updates of the entirety of GNOME in the stable
> updates? Where are the updates to minor feature releases? Oh wait, they
> don't exist! Yet we provide all this for KDE! We even provide a
> semi-official unstable repository (at kde-redhat) with the latest beta KDE
> backported to stable Fedora releases, again where's the equivalent service
> for GNOME?

I think you're using the wrong metric here.

> > and ensuring that they're as involved in the Fedora feature development
> > process as the Gnome developers are.
> 
> Most of those features are upstream GNOME features which just happen to be
> implemented by Fedora developers (or, in most cases more accurately, RH
> developers) and which are inherently or by the nature of their
> implementation GNOME-specific. It's simply unreasonable to expect KDE to
> support every single new GNOME feature at the same time as GNOME, the
> opposite is not true either!

I work on power management. I think this is an important and worthwhile 
feature, and so I spend a lot of time ensuring that Fedora has 
bleeding-edge power management functionality that sets us apart from 
every other OS. This requires desktop integration. KDE does not have 
that level of power management integration. KDE has, in fact, a power 
management UI that commits almost every single power management error 
possible. If KDE is to be considered equivalent to Gnome, then that 
means we can't say "Fedora has awesome power management". Instead, we're 
limited to saying "Fedora (Gnome spin) has awesome power management". 
That's not a useful way to communicate what we're doing.

If there are resources in the Fedora KDE SIG who are willing to help fix 
this, that'd be awesome. It'll involve a fair amount of upstream work, 
including convincing developers who have resisted this issue in the 
past. It's not work I'm willing to do myself.

(I'll be giving a presentation on these issues at the cross desktop 
track at the desktop summit in a week's time. If anyone is there and 
interested in working on this, that'd be great)

> We do track features which are required for distro integration and take
> those seriously. I personally did a lot of the work to make these work.
> Some success stories:
> * ConsoleKit was supported in KDM right from Fedora 7 when it got
> introduced. I did the integration work.

Excellent!

> * PulseAudio got enabled by default in KDE at the same time as in GNOME
> (Fedora 8). Rex Dieter did most of that work (it was coordinated through
> IRC). It involved adding a startup script to kde-settings (no longer needed
> in current releases because PulseAudio's startup is now handled through the
> standard /etc/xdg/autostart mechanism) and making aRts (Fedora 8 was KDE 3)
> work on top of it.

Excellent!

> * FeatureDictionary (i.e. using hunspell throughout) was implemented for KDE
> in the same release which implemented it in GNOME (Fedora 9). I did the
> required patches (backporting the Sonnet Enchant backend to KDE 3's
> KSpell2, adding hunspell support to the legacy KSpell (KDE 3) / K3Spell
> (KDE 4)).

Excellent! But this is a subset of what was added to Fedora in each of 
these releases, and if we're going to advertise certain important 
functionality as a Fedora feature then it needs to be included in the 
Fedora that we encourage users to use. If you want KDE to be considered 
on parity then either KDE needs to implement everything we define as a 
Fedora feature or we need to alter the fedora feature process in such a 
way that features are flagged for the desktops that implement them. If 
that's what you want, then make that argument rather than complaining 
about the naming. You're trying to put the cart before the horse.

> > Once that's done then it's time for a discussion of how the options should
> > be presented, but right now claiming that the two are equivalent is simply
> > false. Changing the text on the website doesn't alter that. Fix reality
> > before trying to fix our description of it.
> 
> "Reality" doesn't need fixing, the website does. The work is being done
> already.

I disagree, but I think that's pretty clear by now.

> > (And if KDE developers are failing to get involved in Fedora because of
> > the layout of the download page, then I think there are larger
> > problems...)
> 
> Can't you understand that volunteer developers are reluctant to work on a
> distribution which considers their work second class?

Can't you understand that a distribution will not consider a desktop 
first class until it has feature parity?

> You seem to either be completely unfamiliar with or have completely lost
> touch with the concept of volunteer development, judging both from your
> repeated insistence on full-time developers and from your lack of
> understanding of what motivates or demotivates volunteers. Getting
> paychecks from RH is not the only thing which can motivate people to
> participate in Fedora!

I agree. The multiple years of unpaid work I spent on Debian and Ubuntu 
ought to demonstrate that. I care enough about Fedora that I spend a 
significant amount of time working on it outside my paid hours. Many of 
the contributions I've made to Fedora are entirely out of the scope of 
my job, but I do it because I care about producing an OS that's 
competitive with anything else on the market. 

> I do a lot of work for Fedora completely for free, because I like Fedora and
> want it to be great. Getting continuously criticized along with other
> fellow SIG members (most of which are also unpaid volunteers) for
> supposedly not doing enough when we're even doing things the GNOME
> maintainers are not doing (such as new version backports, both stable ones
> in official updates and unstable or regression-prone ones in a separate
> unstable repository) is extremely frustrating and demotivating.

I've never claimed that you're not doing enough. The assertions I've 
made are:

1) More Fedora developers work on Gnome than KDE
2) As a result, KDE lags behind Gnome in implementing Fedora features
3) This justifies discriminating between the two desktops 

You're arguing that changing (3) will result in (1) and (2) changing. 
I'm arguing that (3) isn't going to change until (1) and (2) do. That's 
not a criticism of the KDE SIG or their work. I'm genuinely sorry if you 
interpret it as such.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org