"Thorsten Froehlich" <Tho### [at] csicom> writes:

> >BUT take a look at http://www.ssc.com/linux/Eric/cathedral-paper.html !
> >It does make sense to publish half finished, deficient code. If it works
> >for a big project like Linux, why not for POV-Ray? I really don't understand
> >why the povteam doesn't take advantage of the big number of povray users
> >with programming skills. 
> >
> >As mentioned in the article there should always be two versions: a stable
> >one and an experimental one. In the case of POV-Ray this is currently
> >version 3.02 and 3.1b. Since the source code is availible for final versions
> >it is already possible to screw it up and provide it to the net. Why are
> >the povteam members so frighten to show a partly finished product and
> >use the work of others that results from this publication? In my honest
> >opinion POV-Ray is a perfect candidate for a bazaar style development.
> >The povteam would still be the source of stable versions and the central
> >point for exchange - just like Netscape for the Navigator.
> 
> Speaking *only* for myself, I don't want to see CD distributors make 
> profit from the work I have done. 
[...]

No, no. This wasn't the point. I do not want you to change your policy
about distributing POV-Ray. POV-Ray is small enough to be distributed
via the internet and in a single package. (However, POV-Ray is
currently included in at least one distribution of Linux and therefore
sold with it.)

I'm sorry for my bad English, but I hope that you will understand my
suggestion this time:
What I wanted to say is that your group should change your attitude 
about including fixes and improvements from people that are not part
of your team. In fact, in mu opinion it seams advisable to even encourage 
programmers to find and report bugs or add missing features.

> And it is still a lot of work just for Eduard, Anton and me to keep out 
> Macintosh sources in sync...imagine 20 people working on it. POV-Ray is a 
> "small" project compared to Linux, it would be very difficult to direct 
> work for more than 100 people!

Well, if it is possible for Linux, why not for POV-Ray. You even say, 
that it is a "small" prject compared to Linux.

Let me provide a concrete example: Jochen Lippert has added a nice new
object type named "sphere sweeps" to POV-Ray.
(http://www.informatik.uni-bremen.de/~vader/pss/pss.html)
How about the chances that this code will be included in future releases
of POV-Ray? Of course some testing has to be done before it can be
included, but it was quite easy to include his files in my custom version
of POV-Ray on a Unix system. I guess that the same is true for Windows,
even though Jochen wrote his code on a Mac.

Or what about the "Isosurface Patch". I think that these extensions are
of great value for the users. I know, that there are at least two errors
in this patch, but will this code ever go into the official distribution 
once it is cleaned up? My impression is that the answer is "no", and it 
makes me sad. Supposing that the authors of this patch agree, the
inclusion of (parts of) the patch would improve POV-Ray -- with less
effort than programming the same features by the povteam.

I've written a new nice recursive pattern for POV-Ray. Of course it's
not as much of work as the previous examples. I won't clean up the
code, because I'm almost sure that no one will compile a separate version
just for this feature and on the other hand it will never be included
in the official release.

Apart from new features it even strikes me more that detailed bug reports 
aren't encouraged by releasing the source code. A short message before
and/or after each rendering can warn the user about the beta version.
Are you really afraid, that somebody will remove this message and
distribute the beta as a final version?

Once, I have written a bug report including a hint what to change in the
source code to Chris Young. I've never got a reply and the bug is still
included in the current beta. Why does this happen? This really discourages
me to look for bugs in future. I guess that the same is true for other 
people with some interest and knowledge in programming, maths, and raytracing.

Best regards,

Thomas

-- 
Tho### [at] uni-konstanzde
http://www.informatik.uni-konstanz.de/~willhalm/
Tschieses lavs ju
An den BND: Eisbombe Fremdenfuehrer Siegfried Heilsarmee

Post a reply to this message