Ron Parker wrote:

> 
> Actually, there's another reason the team can't just get someone to build
> a SunOS version and send it to them: it might not be just a straightforward
> compile.  There might be problems with the port that would need to be
> resolved by someone who understands the OS it's running on.

I can agree with that point.  You must be speaking about future versions because
the current versions that I have ( 3.1e & 3.1g ) are no problem.  I imagine, I
hope, that future versions will take advantage of multi-processor systems and
thus there may be some interesting problems with a build.  Perhaps.

> I seem to
> recall that the team was looking for a member who could do SunOS work
> about a year ago, long before I was asked to join.  The lack of a current
> SunOS version is not for lack of _wanting_ one, it's for lack of someone
> who was willing to commit to the amount of very real work that's required
> to build a working version and stand behind it.

Well, the current versions of 3.1e and 3.1g work fine.  At least they seem to. 
I don't have any benchmark test to compare against.  I assume that if a bench
mark image was produced with no AA and no jitter then it would be possible to
test for a standard output for a given input.  With the absence of a calibration
reference, I can assume that my versions work well.  Would I stand behind them? 
No.  I have a fully booked life and my health will not allow another task.  Even
if it would be one that would be a passion. 

> 
> >Again, ask the question.  What difference does it make if you support the build on
Solaris
> >at all.  It is NOT supported now.  There has not been an official release of
POV-Ray for
> >Sun Solaris in how long?  A year?  Two?  The current version for SunOS or UNIX on
the
> >"Official" POV-Ray web site is so out of date that I am surprised that it is there
at all.
> 
> There are still 3.0 and earlier versions for other operating systems too.  Just
> because they're obsolete doesn't mean they're any less official.

I don't know if official matters anyway.  If a user could get a ready to go
tar-ball that unpacked into /usr/local and was ready to run, they would be
happy. 

> 
> >This is lip-service.  There is no need to buy a HP9000 system to build POV-Ray for
HPUX.
> >There is no need to buy a Sun Ultra 2300 for UltraSparc and Solaris 8.  (The
pre-release
> >for Solaris 8 is out)  Look at the size and number of the POV-Ray user community. 
I would
> >hazard to guess that most POV-Ray users are serious computer professionals.  How do
you
> >think that Linux bacame what it is today?  By allowing many people to participate
in the
> >open source concept and by using the talents of your vast user community.
> 
> True, to a point.  If you want to build the current source and contribute any
> source patches back to the team, there's nothing keeping you from doing that.
> There's also nothing keeping you from compiling a custom version and
> distributing it.  But I hope you understand how someone wouldn't want to stand
> behind and support a distribution that they couldn't even run, let alone test.
> 

I have had the thought.  I have read through some of the source and I'm not
clear on how all the modules work yet.  I don't know if I'd ever be.  I'd love
to make a project out of a complete rewrite that was multi-threaded and network
based but I have to pay for the house first.  :)

> 
> >If anyone wants my build, they can send an email and ask for it.  No problem.
> 
> If it's not too big, I'd be happy to host a binary on my website, such as it
> is.

I think I should just post my kit on a web page somewhere.  I was worried about
the transfer rate but since I have no idea what the demand will be, I should
just post it and see.  I was taking an inventory here and I find that I can
probaby build a single tar-ball that would have binaries for Solaris 2.5.1 (
x86, Sparc, UltraSparc ) and Solaris 7 ( Sparc & UltraSparc ).  At the moment I
am running on a Sun Ultra 10 and I have both the 3.1e and 3.1g binaries in
/usr/local/povray31.  I don't know why I keep both.  I was concerned that 3.1g
may have problems so I move slowly from version to version.  In any case, I'll
figure out how big an UltraSparc, Sparc, and x86 install kit would be.

> I'm sure someone dedicated could sort out the resulting minor legal
> mess and carry on the work.  But there's a kernel of truth to what you say: as
> I understand it, the reason there's not a 3.1 SunOS build is that the person on
> the team who was doing SunOS left to pursue other interests.  Last I heard, the
> position was still open.

Hmmmm.  I am wondering just how much work is involved.  As for dedication, well,
how many people do you know that have bought Sun hardware ( Sparc & UltraSparc )
for home use?  Very few because its damn expensive and you have to climb up that
UNIX learning curve.  Its just so much easier to click and point mindlessly with
Windows.  But then again, RedHat Linux and Caldera are looking really nice
lately.  And what about the recent purchase of StarDivision by Sun
MicroSystems?  Must be a trend here.  There may be a valid reason to support the
new 64bit UNIX hardware that will be usable by everyone.  But I wonder how much
work is involved.

> 
> Also, I get the impression that you think I'm speaking for the team here.  While
> I'm a member of the team, I don't presume to speak for them.  That's Chris
> Young's job.  I'm just stating my opinions, which would be roughly the same even
> if I weren't a member of the team.

That sounded like a disclaimer.  In all matters, we only speak for ourselves. 
We sometimes have the illusion of a group behind us, but they are usually behind
us, not beside us.

Dennis Clarke
dcl### [at] interlogcom

Post a reply to this message