Abstracting algorithms (graph depth first search)

[Paul Moore]

"Raymond Hettinger" <vze4rx4y at verizon.net> writes:

> I took the class based approach and it is just a fast as a toplevel
> function.  Users subclass the solver and supply ann iterator or
> generator for the edges from each node.

That's good to know. Instinctively, it felt like it would be slow, but
instinct isn't a particularly good guide here...

> The generic solver is about 25 lines.

That's the frustrating thing - writing this stuff really is easy in
Python. The problem comes in remembering the details of the
algorithms. I can write 25 lines of Python no problem. 25 *working*
lines seem to elude me, though :-) I'm still working through the
options and trying to decide whether the extra 25 lines of
infrastructure are worth the time saved making sure the algorithm's
correct :-)

But I was definitely right in one thing - graph algorithms are complex
enough to be interesting, without being too specialised to be worth
bothering with unless you have a specific need.

[Just wrote a Python version of some C++ graph code, which took a
fraction of the space, and was far more readable. Thanks, Guido :-)]

Paul.
-- 
This signature intentionally left blank