Slow Python - what can be done?

[Jason Sewall]

I was out of town for a few days and thus had to rely on Google's only 
okay support for news. When I get a moment, I'll be working on a 
real-deal version of the algorithm and worry about optimization.

I understand your suggestion about the complex number. It's probably a 
good idea and it is reminiscent of quaternions (although in the plane).

I'll let you know,

Jason

Peter Otten wrote:
> Jason Sewall wrote:
> 
> 
>>It's affine for the one-line-pair case, but the full algorithm actually
>>uses n line pairs to describe a nonlinear transformation. Since the
> 
> 
> Would you care to post a small driver program (if feasible)? I'd like to try
> the "complex numbers as points" approach on it.
> 
> 
>>transformations are weighted for each pair, the matrix solution doesn't
>>cut it for the general case.
>>
>>Of course, since I didn't mention it, you wouldn't have known that -
>>sorry.
> 
> 
> No problem.
> 
> Peter
> 
>