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

List:       freedesktop-xorg
Subject:    Re: Xegl lives!
From:       Jon Smirl <jonsmirl () gmail ! com>
Date:       2005-05-26 16:59:17
Message-ID: 9e47339105052609597299eb7e () mail ! gmail ! com
[Download RAW message or body]

On 5/26/05, David Reveman <davidr@novell.com> wrote:
> I think it's much better to create an API design for an imaging model
> that is hard to accelerate right now but what application developers
> want for the future than to create an API that is well accelerated today
> but uninteresting to application developers in a year or so.

I think this logic is backwards. There are only a few algorithms that
can be efficiently accelerated in hardware. And all of the hardware
we're likely to get implements those few algorithms. If instead we
pick an imaging model that doesn't map well onto these algorithms
we're going to be stuck with non-accelerated drawing forever.  D3D and
OpenGL are both designed around the algorithms that can be
accelerated.

It's not clear to me that designing an easy to program to but slow API
is a better choice that a harder to program to API that is 50x faster.

An example way of making something like Cairo completely
unacceleratable would be to specify at the pixel level the line
drawing or polyfill algorithms.

--=20
Jon Smirl
jonsmirl@gmail.com

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