Re: [rp-ml] STL files

Adrian Bowyer wrote:

> This was a problem with a known simple solution when the STL format was
> defined. As that solution had been around since the time of Poincaré
> (1854 - 1912), if not Euler (1707 - 1783), that is a bit annoying. It's
> the sort of thing that happens when you let mechanical engineers (like
> me) define a data structure...

The rules for determining that a surface is 'clean' are certainly well
understood - and somewhere between your mouse and your RP machine
something should be checking that the model follows those rules.

But I fail to understand how the file format representation impacts
that. It would be hard to come up with a file format that made it
impossible to represent illegal shapes that could still be written
out by existing CAD and 3D design software.

Of course we should be careful not to imply that satisfying the
Poincare/Euler rules solve all the problems. There is still no
similarly easy way to detect self-intersections and such. Toss
a Klein bottle or a Boy's surface into many of those tools and
they say "OK!" because they are not concerned with the direction
of edge connectivity as you'd hope they would be.
Received on Thu Dec 07 08:07:02 2006