Wednesday, September 20, 2006

Overflow

When something happens one can assume it is magick and try random modifications to fix it. Better way is to find what is so specific about the situation where problem happens - and check the edge conditions.

Often the two approaches get combined in one - desperately do random changes, and explain why it worked from the materialistic standpoint (once it worked, that's it).

In this particular case, when normal is exactly aligned with X, Y or Z, char gets overflowed and the normal starts pointing the other way. Stupid.

No comments: