[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: plans for a free, OpenGL-based equivalent to Direct2D ?
From:       "Aaron J. Seigo" <aseigo () kde ! org>
Date:       2010-12-14 21:36:32
Message-ID: 201012141336.41193.aseigo () kde ! org
[Download RAW message or body]

[Attachment #2 (multipart/signed)]


On Tuesday, December 14, 2010, Thiago Macieira wrote:
> Em Terça-feira, 14 de Dezembro de 2010, às 11:17:06, Aaron J. Seigo 
escreveu:
> > having looked a bit at the QML Scene Graph, i really do think it would be
> > tremendously great to work towards making it a bit more generic. it
> > wouldn't  need a perfectly stable API/ABI at first, imho, as it could
> > very easily be built right into each app that needs it for the time
> > being. what's important is the time savings for the developer and the
> > correctness of the OpenGL handling; QML SG does both very well already.
> 
> Why not use one of the existing Scene Graph projects out there then? Ogre?

yes, it's possible to use a variety of different scene graph implementations. 
it'd be nice to have a consistent one, however, especially for content that is 
expected to be used with each other. e.g. HTML-in-QtQuick (which would mean 
QtWebKit-in-QML; if Gecko used the same scene graph implementaton and could 
use a pre-existing scene graph handed to it, we could easily make that Gecko-
in-QML as well). at that point, we could, if done well, share one scene graph 
between QtWebKit-in-a-QML and the QML itself.

there's also something to be said about "optimizing things in one place", 
"avoiding having to optimize the drivers for the needs of different usage 
patterns due to different scene graphs" and "being able to transfer knowledge 
between projects by using the same scene graph as much as possible".

we could really do amazing things if we could even agree on a set of shared 
shaders for elements common to web and qml (which these days is pretty much 
all of them). the combination of predictable state + known shaders == graphics 
driver optimization nirvana.

variance in state (e.g. scene graph strategy) and/or shader (usually 
application specific) just makes it all a mess for the driver writers working 
on optimization.

since QML and HTML5 (and likely Calligra, though i have less insight into 
their specific use cases, so could be wrong there) are going to have extremely 
similar requirements, i bet the bulk of shaders could be shared.

from a platform developer's POV, having such a consistency would be a desired 
goal. it would elevate the ROI for optimization in the driver way up for them.

-- 
Aaron J. Seigo
humru othro a kohnu se
GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA  EE75 D6B7 2EB1 A7F1 DB43

KDE core developer sponsored by Qt Development Frameworks

["signature.asc" (application/pgp-signature)]

>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<


[prev in list] [next in list] [prev in thread] [next in thread]