Hello guys,

  I'm just a poor Linux user who wishes he could use the brand-new and
(apparently) greatly improved versions of POV-Ray. The problem is, the
POV-Team seems to have a bit changed its opinion of porting POV-Ray:
it seems that the panel of officially supported platforms tends to
converge on the single Windows 95 version (maybe I'm a bit
exaggerating here), and the announces they make do not even have a
[WIN] flag. I dont' blame it, I just regret it: they seem to spend a
lot of time on the editor and the interface, which concerns only one
platform, and they feel continuously bugged by people like me who
just require the renderer and no interface (no use for an interface
when you have Emacs).

  My point is, I'm more and more seriously thinking about how hard it
is to make a portable GUI, and I am more and more seriously thinking I
won't resist to try to make one soon. "This guy's just mad, GUI cannot
be portable!" Yeah, maybe. I still want to give it a try. There are
many ways to do it, and if you consider me crazy I would like to know
why, because if not I am really going to try hard. Please give me your
opinion about the following ideas:

 - A Tk interface. I don't know Tk yet, but I'll have to learn it
anyway for a professionnal job, and from what I read it seems quite
easy and quite portable. The underlying language would be either
Python or Tcl, which I also plan to learn in the next few weeks. My
proposed plan would be to extend these languages with C-written
functions that I would get from POV-Ray.

 - An Emacs interface. Would be a bit messy to do, but it would
provide the greatest editor ever (if people really do edit their POV
files with vi, please contact either me or your local analyst), thus
saving time for the POV-Team and providing a portable (and highly
customizable) editor. It could even be done with XEmacs, which would
give the nice graphical elements. I know there's already a major mode
for editing POV files, I could adapt it and extend it to match the new
syntax elements, and add the POV-Ray front-end on top of it.

  I know these methods could not really be as fast as a C-written
interface (despite the difference would not be *that* visible), but
anyway who cares? The hard point in raytracing is number-crunching,
not text editing... And embedding a C function into a Python program
won't slow it down, neither will an external program (the render-only
part of POV-Ray) be slowed down if called from an Emacs.

  Yes, that's a lot of work, but I'm ready to get involved in
it. Anyway I will, but even more if I get feedback about these
methods. Who's interested? Who's interested in the result? Who's
interested in helping (merely by beta-testing, but why not by coding
too)? I need your advice...

  And it can also lead to a bazaar model of development for POV-Ray,
like the one used for many other open source software pieces, which
gives pretty good results (no, my father's name is not Eric Raymond, I
just read his article some days ago). Anyway, let the POV-Team
choose...

  May our rays be forever traced,

Roland.
-- 
Les francophones m'appellent Roland Mas,
English speakers call me Rowlannd' Mass,
Nihongode hanasu hitoha [Lolando Masu] to iimasu.
Choisissez ! Take your pick ! Erande kudasai !

Post a reply to this message