Hi Paul,
I've been following this erudite discussion with interest and a modest degree of understanding.
But one thing seems clear to me - that any quality implementation is going to require at the very least 'big' ints and 'big' floats, and probably exact int and exact floats.
Yes, at the very least.
It would seem that we need tried and tested Boost license implementations - preferably before starting work on a complex geometry problem. No solution can be considered for Boost if it uses any restrictive-licensed components.
I thought about this as well. I think we could really appreciate if Boost could finally get a big int/float number type. There even is partial work lurking on the vault for so long I lost count. OTOH, I don't think that a library CAN'T push that into user space, even if in practice it *requires* users to use a non-boost licensed implementation, provided it is LPGLed or some other reasonably free enough. At the very least, I think gmp, which is LPGL and "needed" by GTL (for all practical purposes), deserves a special consideration. It is a de-facto standard to such an extent that I even question if it is really worth the effort of preproducing that. The same goes for mpfr in the floating point arena. Best -- Fernando Cacciola SciSoft Consulting, Founder http://www.scisoft-consulting.com