Yoooooooooooooo !! > Well, whether that's sufficient is between you and the reviewer of the code.
Oh. I see. > Perhaps the docstring wouldn't have to contain it, but don't you think it should be at least in principle straightforward to document the implementation as well as what your routine tries to do? Yes, of course ! That doesn't scare me, though. > Your original question was along the lines of "I have a description of a computational procedure that when executed literally would immediately lead to infinite recursion. How do I resolve that?". > > One solution, the one you describe, is to realize that the infinite recursion comes from going in circles in a search strategy, so to avoid that you would keep track of where you've been already. What's left to prove is that the rest of the paths you take make up only a finite search space (i.e., your parameters are eventually decreasing). Yup. > Another solution, which I think is clearer, is changing your search strategy to naturally not go in circles. That's how you'd have to describe it if you want it to get published in a somewhat respectable journal anyway, because you'd have to prove your algorihm actually finishes in finite time. I try to avoid respectability at all costs, but I see what you mean. I'd like to keep the code easy to understand, though. That's not always a hard constraint for published papers, even in respectable journals :-P > Splitting up your search strategy in smaller, well-defined blocks also has the advantage that it's easier to reconfigure your strategy later on, should other (better?) constructions become available. And it sounds like this area could still be in flux. Well, splitting sounds nice indeed but what exactly do you point at ? You can't swear that a TD will be built without ever requiring a MOLS to be built at some recursion step. Nathann -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
