The only issue I've had using pure RB code to call OpenGL is with
calling non-opengl1.1 functions and extensions on Windows due to the
need to be able to call functions by pointer - a glaring omission in
RB IMHO, especially now that we can generate function pointers using
AddressOf. Still, I prefer to restrict my plugin use to just this one
requirement, and remain hopeful that one day I can eliminate this also.
True you cannot use GLUT this way, but GLUT doesn't play very nice
with RB anyway, seting up its own event loop and performing its own
windowing, which RB does perfectly satisfactorily on its own. It used
to be the case that RB event loops were an issue, but the new threads
in RB2005 seem superb for use with games, knocking the need for
timers or single-threaded tight loops on the head as far as I'm
concerned. But I digress.
I am sure you are better aware of your project's needs that I am, but
speaking for myself I would be loathe to base a commercial product on
top of someone else's work, even if it was of high quality and I was
reasonably sure of its continued support. Many a debate has raged on
these lists about how tiresome it is to be at the mercy of RS's bug
fixing schedule,and even stellar products such as the Monkeybread
suite are buggy in places - the more different non-open-source third
party products you base you product on, the more fragile it becomes.
Given the relative hardship of porting a plugin to XCode, and then
maintaining parallel projects in XCode and Codewarrior (and possibly
Visual Studio as well, not to mention whatever Linux IDE is
required), I have little confidence that any but the most dedicated
of plugin authors are going to continue full multi-platform support
for their products after the coming transition, (at least until a
better development solution presents itself).
As I mentioned before, it's the x86 compatibility issue in particular
that has given me pause just recently. While I am reasonably
confident that RS will cope with this transition smoothly, I would be
very hesitant to plug anything into my projects right now that relies
on closed-source prebuilt PPC binaries supplied by another individual.
Nick
On 9 Dec 2005, at 01:01, John Balestrieri wrote:
Better access to certain OpenGL components and extensions. I'm not
using it atm, but GLUT is inaccessible from declares, iirc.
Also, to do some optimized stuff for my 3d engine, turning some
classes into plugins will allow for better memory performance.
Time. Before I could compete my gl-headers-to-rb-declares project,
a complete implementation was dropped in my lap, and it worked on
Windows, by Jon Johnson. It was an easy switch to use the plugin,
and would be easy to not use the plugin if I had to (but I am
pleased with it). By not having to maintain an opengl
implementation, I have been able to focus on the top layer
functionality of SSS, etc.
John
On Dec 8, 2005, at 6:07 PM, Nick Lockwood wrote:
I was actually referring to Marco's efforts, but I suppose it
applies equally to your decision.
Can I enquire why it is that you are making this switch? What
benefit do you believe it will yield?
I would have thought that the effort overhead of maintaining at
least 2 and possibly 3 or 4 different plugin components (which, at
present must be maintained in at least two different IDE
environments) would outweigh most potential benefits of plugins,
and I certainly can't believe speed is a factor since all the hard
work is done by external library functions anyway.
Nick
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
|