Also, I feel there should be an easy-to-find redirect function in the symmetric function code (IIRC in sfa.py would be the most natural spot), which would tell you to look in the corresponding polynomial code. My natural place to put such checks would be in the polynomial code given that it could take advantage of the specialized implements as Vincent said.
Best, Travis On Friday, April 24, 2020 at 3:08:38 AM UTC+10, vdelecroix wrote: > > Le 23/04/2020 à 18:53, Nils Bruin a écrit : > > On Thursday, April 23, 2020 at 3:04:46 AM UTC-7, vdelecroix wrote: > >> > >> > >> I think that it would be good to implement this directly of the level > of > >> multivariate polynomials. I opened > >> > > > > What would the uptake of such a method be? If I needed to test if a > > polynomial is symmetric, I would probably end up rolling my own because: > > - I wouldn't know or expect such a method to exist on symmetric > polynomials > > - I wouldn't know what the particular conditions and assumptions on > the > > input would be (and whether the author silently had different > definitions > > in mind than I had). > > Because it's so easy to program myself, I would probably decide that's > > faster than figuring out if the library version is appropriate (if I > knew > > if it existed in the first place). > > > > In terms of use: I think it's very rare in a computational setting to > want > > to know if a polynomial is symmetric and not want it expressed in > > (elementary) symmetric polynomials. That's a routine I'd look for -- > > although surprisingly, this is something that routinely gets done poorly > in > > computer algebra systems (including magma!), so I'd probably end up > > implementing elimination anyway. > > > > (in this case, it's clear the code is needed in at least one place of > the > > library, so it's not really a loss to put it somewhere, but for code > > maintainability, I'd keep it closer to where it's used and then perhaps > > refactor if it shows the demand is there to have it in a more accessible > > spot) > > I agree that if you have a symmetric polynomial at hand you most > probably don't want it expanded (or you are probably doing something > wrong). > > But first of all, to make it fast there is no other option than > making it at the level of polynomials themselves (and even then > the generic version it is quite slow because I believe conversion > back and forth between Singular and Sage). Secondly, as a developer, > if I am looking whether such feature is implemented I would never > look into sage/combinat/sf/monomial.py. To my mind, there is more > risk of duplication by implementing it close to its usage rather > than close to the data structure it concerns. And finally, I > implemented the latter in ticket #29553. > -- 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 view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/e0ad958c-4462-4e75-83aa-61bd8f24bd8d%40googlegroups.com.
