Well, I just missed the deadline - when I started to type this, there were 
22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's 
not going to make it.

Rats.  I'm sooo frustrated.

Well, there were a few errors in the image anyway, my radiosity settings 
interferred with my terrain and washed it out so the impact wouldn't have 
been there.  Same terrain I've been using for 3 weeks, but I fiddled with 
the texture in the last 3 days and apparently it didn't work with radiosity.

I think I ran into a bug with fog 2 that's different than the known bug, 
I'll try and duplicate it.  Probably I messed up the SDL.

The worst, however, was two multiple crashes of MegaPOV on the scene at 
about halfway through, this afternoon.  After attempting to continue with 
POVRay, it worked, but scrambled the top half of the image, and I was scared 
to mess with it.  After looking at it just now, it appears the file saved 
was correct, so it must be a display bug with +C switching between MegaPOV 
and POVRay.  I'll post the crash address if I can find it.  I was SO happy 
yesterday as I went to sleep and started the final image ...  well, clearly 
I should have started earlier.

I had a few test runs that weren't too bad (not as much detail) but I hadn't 
saved the detail zoom in so I would gather I can't post them anyway.  Next 
time, I'll do complete dry run before the final image 'just in case'.  If 
there is a next time ...

oh yea, 4 mins to go and line 456 out of 571 ... there it goes, 00:01.  :(

        == John ==

Attachments:
Download 'MegaPOVCrash.txt.dat' (17 KB)

> Well, I just missed the deadline - when I started to type this, there were
> 22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's
> not going to make it.


That sucks =(

But hey, finish the image, and show us the final product in
povray.binaries.images. It's not like it was a waste of time.

> After attempting to continue with
> POVRay, it worked, but scrambled the top half of the image, and I was
scared
> to mess with it.  After looking at it just now, it appears the file saved
> was correct, so it must be a display bug with +C switching between MegaPOV
> and POVRay.

I've seen that without using MegaPOV. In fact, I see it every time I use +C.
File outputs alright so I don't really mind it.

 - Slime
 [ http://www.slimeland.com/ ]

"John D. Gwinner" <joh### [at] cornelledu> wrote in message
news:41ff37a1$1@news.povray.org...
> Well, I just missed the deadline - when I started to type this, there were
> 22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's
> not going to make it.

Damn, that's a real shame John!  I was really hoping no one would run out of
time :/

Still, like Slime said, it's not a total loss - at least you have another
render under your belt.

I know how it feels though, to miss a deadline like that... :((

Lance.

thezone - thezone.firewave.com.au
thehandle - www.thehandle.com

"John D. Gwinner" <joh### [at] cornelledu> a écrit dans le message de
news:41ff37a1$1@news.povray.org...
> Well, I just missed the deadline - when I started to type this, there were
> 22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's
> not going to make it.
>
> Rats.  I'm sooo frustrated.
Too bad! Must be a heartbreaker :(((


Marc

http://citeseer.ist.psu.edu/108629.html

If you managed to get it finished, why don't you submit it anyhow ? If
you email the support line they can re-enable your login. Then, at least
it will be displayed with the other entries.

-- Chris

John D. Gwinner wrote:
> Well, I just missed the deadline - when I started to type this, there were 
> 22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's 
> not going to make it.
> 
> Rats.  I'm sooo frustrated.
> 
> Well, there were a few errors in the image anyway, my radiosity settings 
> interferred with my terrain and washed it out so the impact wouldn't have 
> been there.  Same terrain I've been using for 3 weeks, but I fiddled with 
> the texture in the last 3 days and apparently it didn't work with radiosity.
> 
> I think I ran into a bug with fog 2 that's different than the known bug, 
> I'll try and duplicate it.  Probably I messed up the SDL.
> 
> The worst, however, was two multiple crashes of MegaPOV on the scene at 
> about halfway through, this afternoon.  After attempting to continue with 
> POVRay, it worked, but scrambled the top half of the image, and I was scared 
> to mess with it.  After looking at it just now, it appears the file saved 
> was correct, so it must be a display bug with +C switching between MegaPOV 
> and POVRay.  I'll post the crash address if I can find it.  I was SO happy 
> yesterday as I went to sleep and started the final image ...  well, clearly 
> I should have started earlier.
> 
> I had a few test runs that weren't too bad (not as much detail) but I hadn't 
> saved the detail zoom in so I would gather I can't post them anyway.  Next 
> time, I'll do complete dry run before the final image 'just in case'.  If 
> there is a next time ...
> 
> oh yea, 4 mins to go and line 456 out of 571 ... there it goes, 00:01.  :(
> 
>         == John == 
> 
> 
> 
> ÿþ<
I commiserate.

John D. Gwinner wrote:

> Well, I just missed the deadline - when I started to type this, there were
> 22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's
> not going to make it.

The lesson is everyone attempting to enter the competition should have
rendered a submission quality image each weekend, from the start of the
contest until now. Then if you can't render your best version, you can
submit the second best.

Frequent releases also teach how long a render takes.

This is a "best practice" in software engineering.

-- 
  Phlip
  http://industrialxp.org/community/bin/view/Main/TestFirstUserInterfaces

*lol*

Of course.

"emkaah" <emkaah@yahoodotcom> wrote in message 
news:41ff60f2$1@news.povray.org...
> http://citeseer.ist.psu.edu/108629.html
>

Phlip wrote:
> John D. Gwinner wrote:
> 
> 
>>Well, I just missed the deadline - when I started to type this, there were
>>22 mins to go or so, but at line 440 out of 571 to go ... obviously, it's
>>not going to make it.
> 
> 
> The lesson is everyone attempting to enter the competition should have
> rendered a submission quality image each weekend, from the start of the
> contest until now. Then if you can't render your best version, you can
> submit the second best.
> 
> Frequent releases also teach how long a render takes.
> 
> This is a "best practice" in software engineering.
> 
Yeah yeah yeah, but the real world is messy.  I did a lot of due 
diligence full blown testing myself, along with clever speed testing, 
and built a lot into my code to facilitate that,... but I still got a 
big surprise during the final push at the end.  I, myself, was forced to 
enter a set of renders that I did in just the spirit you describe.  But 
they look embarassing to me now. They could have been improved 
dramatically with just a few "plan B" tweaks, once it was obvious that a 
technique I was depending on wouldn't work, but instead I had to watch, 
helplessly, as the clock ticked down, just as John did.

Chris Cason wrote:
> If you managed to get it finished, why don't you submit it anyhow ? If
> you email the support line they can re-enable your login. Then, at least
> it will be displayed with the other entries.
> 
> -- Chris
That's a nice invitation.  After all, one of the main attractions is the 
increased level of exposure, as much as the prizes.

"John D. Gwinner" <joh### [at] cornelledu> wrote:
> The worst, however, was two multiple crashes of MegaPOV on the scene at
> about halfway through,
?

have you tried a ram checker (memtest86)?
i found out (the hard way) that foulty ram only shows up in large renders.
(i got errors like "triangle misspelled" in the midle of a large (auto
generated) file with thowsands of triangles.. (and hang-ups :-) )

jaap.

I just barely managed to submit my entry in time.  I am by no means a
professional raytracer, but I had to challenge myself to get somewhere.
Just as I finished several hours of rendering, I noticed that one of the
meshes was missing from the foreground.  It was too late to go back and
replace it, and the scene was still intelligible without it.  I promised
myself I'd enter no matter how crappy my submission was.  At least now I
have a *very remote* understanding of some of what goes on underneath the
program's hood.  I think it would be really nice to have some interactive
graphical interface or preview option so I don't waste three hours
rendering the whole scene just to detect a simple problem in one little
corner.
Anyway, I'm sorry you couldn't finish yours in time.  I can't say my chances
are any better having submitted an entry than not, though.

> I think it would be really nice to have some interactive
> graphical interface or preview option so I don't waste three hours
> rendering the whole scene just to detect a simple problem in one little
> corner.


+SP16

 - Slime
 [ http://www.slimeland.com/ ]

> Frequent releases also teach how long a render takes.

Actually, I *was* doing that.  What I wasn't doing was the 2 'detail' 
images, as I had been doing close ups of the detail shots in other renders, 
so I didn't need to preview those.  The net result is that I was well aware 
of how long a render took, but the backup image wouldn't have met the 
requirements.

As I noted in the original, due to a versioning mismatch, the photon times 
suddenly shut up dramatically the last day, but that's an integration issue. 
Like all integration issues, proper planning from the start would have 
corrected that, so of course it's my screwup.  *so* much went wrong at the 
end there was only so much I could do.  Of course, I could have done better.

  The image is part of a series for a book I'm writing so it will live 
again, however.

> This is a "best practice" in software engineering.

  Right.  I've signed checks (authorized funds) for 16 million for software 
projects.  (no kidding).  You?  :)


        == John ==

Thank you (Chris).  Let me see if I can pin down the fog bug, rerender, and 
then I'll try and post something.

I agree about the exposure - this image in particular shows some interesting 
'code' things - I have a city that's built programatically, although 
distance shots show it a bit better.  The ability to easily extend POV Ray 
with SDL is IMHO one of the big attractions, compared to other programs.

        == John ==

"Jim Charter" <jrc### [at] msncom> wrote in message 
news:41ffacc3$1@news.povray.org...
> Chris Cason wrote:
>> If you managed to get it finished, why don't you submit it anyhow ? If
>> you email the support line they can re-enable your login. Then, at least
>> it will be displayed with the other entries.
>>
>> -- Chris
> That's a nice invitation.  After all, one of the main attractions is the 
> increased level of exposure, as much as the prizes.