Xplo Eristotle <xpl### [at] infomagicnet> wrote in
news:3d52cb38@news.povray.org: 

> Warp wrote:
>> Philippe Lhoste <Phi### [at] gmxnet> wrote:
>> 
>>>Now, it would be nice if somebody give a patch to correct some of
>>>these bugs, without affecting the regular behavior (speed,
>>>rendering)... 
>> 
>> 
>>   The example scene I posted does not show any bug. It simply shows
>>   that 
>> floating point calculations have limited accuracy. A bug is a
>> programming mistake, and limited floating point accuracy is not a
>> programming mistake. 
> 
> It *seems* like a bug, though. Would it be possible/wise to make
> POV-Ray report when rounding errors like these occur, so that the user
> won't tear his hair out looking for bugs in POV-Ray or his scene?

That's what I, and the POV team meant, I think.
It seems like a bug, because it doesn't give the expected result.
That doesn't mean there is a way to correct them, only workarounds.

Well, perhaps there is a way, for the glutons for punishment :-)
There are a number of unlimited precision floating point libraries out 
there. So if you don't care for speed and memory consumption, you can 
rewrite POV-Ray to use them, at least for critical sections, instead of 
the standard floating point routines (relying on the fast FPU...).

I doubt it will worth the effort, though.

Regards.

-- 
--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--
Philippe Lhoste (Paris -- France)
Professional programmer and amateur artist
http://jove.prohosting.com/~philho/

Post a reply to this message