On Tue, 9 May 2000, Andy Ross wrote:
> This isn't a vote. I'm not involved enough in the OpenGL "community"
> to really warrant one, but let me make an argument for C:
>
> 1.) Old-style extension usage is broken:
This isn't true. You are assuming that all programs are required to
be portable - and portable at the binary level. All *ABI* programs
should be - but not all legal OpenGL programs under Linux need the
ABI.
In fact there are a vast number of programs that people have written
just for their own use (or the use of a few co-workers) that don't
have to be portable at all. There is also a large body of code that
is distributed in source form (THIS IS LINUX - REMEMBER!) that only
has source-level portability - for which the traditional OpenGL
extension mechanism works just fine.
> 2.) These are INHERENT flaws. They can't be fixed. Old source and
> binaries will be FOREVER plagued by these problems, regardless of
> what the default behavior of gl.h is.
But right here is your problem - you are assuming that a lack of binary
portability is a flaw - when in fact it is not for all but a handful
OpenGL programs written for Linux (Quake, Heretic II, etc).
In terms of the number of people writing code for Linux, only a tiny
proportion will ever care about this ABI. Admittedly those few people
will be making more money than all the rest put together - and certainly
we need to encourage those people...but I don't think you can justify
pissing off the free majority in that cause.
> Stated differently, a
> "compatible" OpenGL Base ABI will be no more compatible than the
> pre-existing environments, which aren't.
Using a token to turn ABI compliance on is all that is needed to make
it arbitarily compatible.
Steve Baker (817)619-2657 (Vox/Vox-Mail)
L3Com/Link Simulation & Training (817)619-2466 (Fax)
Work: [EMAIL PROTECTED] http://www.hti.com
Home: [EMAIL PROTECTED] http://web2.airmail.net/sjbaker1
