On Fri, Mar 13, 2009 at 7:06 PM, Ralf Hemmecke <r...@hemmecke.de> wrote: > >>> Of course, then QQ(1)==ZZ(1) would return false. But I really don't >>> see a problem with that. > >> I would find that super inconvenient. > > Well, maybe later you'll appreciate my suggestion a bit more. > >> How about "parent(a) == parent(b) and a == b" > > (of course "a==b" also must have "parent(a)==parent(b)" inside, because > that is the condition that decides whether a coercion lookup should > start or whether the == just works inside the parent. So > "parent(a)==parent(b)" has to be evaluated twice. > >>> Note that it might even be a speed consideration. Always looking up >>> coercions (even if they are cached) is not a good strategy. And the >>> current '==' not being transitive might also hide bugs that are really >>> hard to detect. > >> People have yet to show an example where this really hides any bugs. > > Good luck for the bug search whenever this will happen.
This is ridiculous. If we made a==b true only if parent(a) == parent(b), we would get nonstop complaints and confusion from users, and would be doing something different and massively more *pedantic* than every other math software system I have ever used. -- William > Actually, programmers need just be a bit careless. They have tested a==b > and b==c and then use c where they should have used a. If the code is > complex enough, you can even read the code and don't spot the error. We > all are so used to the fact that something that looks like == behaves > like an equivalence relation. > > Ralf > > > > -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
