In article <3d0ae109$1@news.povray.org> , "Karl J. Anders" 
<kar### [at] webde> wrote:

> with all due respect to the amount of work the team has put into
> this software ( and is still putting into, of course ) I can't refrain
> from stating that I don't like two changes that have been made:
>
> - removing an operator in release candidate (!) #6 (^)

This has been removed to avoid getting complaints from people who don't
understand that there can be more than one way how something works.  The
feature used to work one way and thus we got complaints from those favoring
the other way of operation.  Changing it would have resulted in the other
group complaining.  Others seemed to complain just for the sake of
complaining and they didn't mind the original (MegaPOV) implementation not
working different.  As the feature is redundant anyway and was only a 3.5
beta addition and was not in 3.1 it has been removed because we don't want
to have to waste our time with endless pointless discussions how it works
and how some people expect it to work.

> - the often discussed problem about vnormalize( <0,0,0> ) = 0
>
> The second issue breaks probably thousands of scenes, so it should be
> at least mentioned in the chapter about "changed features that may break old
> scenes". I think that quite a few people have written ( or are writing right
> now ) macros to reproduce the old behaviour. I propose to add such a macro
> to an appropriate include file ( or even better, reinstate old behaviour ;)

There was no old behavior.  There never was a legal result being returned
for a zero length vector and the documentation clearly stated it.  You
cannot normalize a zero length vector, it is undefined and always was.  The
problem was that the parser didn't catch this...


    Thorsten

____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg

I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message