Hi Mathias,

good to have you back! :)

You wrote on Mon, Jan 12, 2009 at 05:39:49PM +0100:
> Thorsten Behrens wrote:
> > While I definitely applaud all efforts spend improving performance
> > of OOo, I see serious overlap in scope with existing projects. 
> Of course, the overlap is there and that's one reason why we failed to
> make progress on performance: we have overlap but didn't manage to get
> people to work on the overlapping parts. So it's good to have a
> dedicated group of people that gives performance a priority.
>
I'm more than happy that seemingly there'll be people assigned to
this, don't get me wrong (my mail stated exactly that, but it seems
the impression was still a different one).

> In the same way as we have e.g. a graphics or a framework project though
> it overlaps with other projects as well.
>
Coincidentally, there was a recent inquiry about the resource activation 
framework in sd (->graphics), and why it was not part of the framework 
module... ;)

> There's no requirement for the OOo projects to be disjunctive.
>
No, but as mentioned elsewhere, having a suboptimal state is no good 
reason to make it even more suboptimal. And the graphics/gsl/framework/ 
etc. projects at least have separate, physical code modules; whereas a
"performance project" will necessarily work all across the board,
and touch code in literally all modules, by default, and not
exceptionally, because some once-in-a-lifetime drawing layer rework
takes place.

Kind of reminds me of the "virtual UI project" - basically, all I'm
saying is that the very nature of performance optimizations in OOo
does not lend itself to the very special (segregative) semantics of 
CollabNet's projects - even less so than in other cases, if you 
will. ;)

If there's a need for a dedicated mailing list (I doubt it, every dev
should be interested in that topic), revive performa...@tools. The
people running this endeavour will be our heroes anyway, no need to
trap them inside an incubator glass cage. :)

Cheers,

-- Thorsten

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to