Hi Waldek, in general, I am, of course, for generalisations. Do you only want Z --> R:GcdDomain Q --> Fraction(R) ?
What is your concrete application for such a generalisation? Do you have also other things in mind? I used it in my article http://axiom-wiki.newsynthesis.org/DancingSambaRamanujan I haven't checked now, but I have the feeling that I would get the result also with your proposed changes and non-inlined integer operations. However, I am strongly in favour of keeping the domain IntegerLocalizedAtPrime as is (or even optimize it more). The integer case is probably the most used case. I hope that you don't need the generalisation right now. I'm currently working on a generalisation of the samba algorithm. Maybe that requires also some generalisation of some localized domain. I cannot say yet, since it's still work in progress. Hope that helps. Ralf PS: I'm sure, you have also found fricas.github.io/api/Localize.html . That domain looked somehow weird to me. For a localization, I would expect a ring R and a multiplicative closed set S and then form S^(-1)R. Why Localize(M, R) wants R to be a CommutativeRing, is not completely clear to me. Maybe only, because M should be a Module(R). On 05/10/2018 04:30 PM, Waldek Hebisch wrote: > Various localizations seem to play important role in algebra > so we would like to support them. ATM I am not sure how > to implement more general localizations. However, as > a small step I looked at IntegerLocalizedAtPrime. It compiles > when I replaced Z by a GcdDomain R and Q by Fraction(R). > And AFACS it should work correctly with such assumptions. > The only exception is call to 'prime?' which checks is > argument is indeed a prime. > > Ralf, is there some strong reason to keep this domain > limited to integers? General version probably will > be slower because calls to integer operations can be > inlined and general R would use normal calls -- if > integer case is critical we can have conditional > implementation so that speed in special case is > better than in general case. -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To unsubscribe from this group and stop receiving emails from it, send an email to fricas-devel+unsubscr...@googlegroups.com. To post to this group, send email to fricas-devel@googlegroups.com. Visit this group at https://groups.google.com/group/fricas-devel. For more options, visit https://groups.google.com/d/optout.
