Thanks for the quick reply Bill. Your comments make more sense (to me) now, since I see that your concerns are more about the way categories are handled in Sage, or at least how it interacts with the parent/element system.
At this stage there is no category theory in Nemo (it wouldn't be appropriate in Julia itself, since that is a general purpose language, not really a CAS). There is a Nemo development list by the way. It's nemo-devel on Google Groups. It's public, so anyone can request to join. If you prefer to discuss it further there, you are most welcome. Though I must warn that Nemo is a very new system and there are no efforts started or planned for category theory. Nemo isn't trying to be the next Sage, so it's not even clear that thinking about categories in Nemo is even meaningful. Nemo exists primarily to do something innovative for that part of computer algebra/number theory that requires very fast generics. That doesn't really cover category theory I don't think. However, our group in Kaiserslautern is good friends with the group in Achen doing HomAlg and Cap, which are category theory projects making use of Gap 4. There may be some interaction between us and them in the future, since at the very least we plan to interface with Gap. Bill. On 1 October 2015 at 18:38, Bill Page <bill.p...@newsynthesis.org> wrote: > [Changed thread subject from: Sources of funding - perhaps computer > manufacturers? ] > > What I find hard to swallow is the peculiar mix of "parent", > "category", and Python data types (class system and inheritance). In > spite of the available documentation in the category system in Sage, I > really don't know what to use when and how to map it properly to > abstract mathematics in such a way as to gain as much as possible from > the existing Sage libraries. The situation is very different in > Axiom/Aldor. > > It is good that you mention Aldor because it was designed specifically > as the "next generation" of Axiom library compiler but for entirely > non-technical reasons never achieved this status. On the other hand > the Aldor documentation is still the best available documentation of > the Axiom type system. > > http://www.aldor.org/docs/aldorug.pdf > > Aldor is currently supported only by the FriCAS fork of Axiom. Aldor > development itself is not dead, merely sleeping, and apparently a low > priority for the several people involved. > > https://github.com/pippijn/aldor > > To the best of my knowledge the Sage-combinat development of the > category system in Sage was motivated largely by the experience of > some of the developers with the Axiom-like category system that was > implemented in MuPad. MuPad was originally modeled after an early > version of Maple and so far as I know has a syntactic/symbolic > orientation at it's core (like Maple, Mathematica and Maxima) rather > than a more "algebraic" one. The MuPad developers borrowed from Axiom > in developing a more static type-oriented architecture for the > library. Maple itself as you might know has taken a rather different > approach. But in any case, compared to Axiom, the MuPad category > systems seems to me to be more or less grafted-on. And the Sage > category system seems much more so. > > The little that I know about the Julia type system, mostly from > discussions about Axiom and Aldor on that mailing list about a year > ago, is that it has a more malleable design that one might expect to > enable a more integrated implementation of the central ideas of the > Axiom category system. So far as I know however no one is working on > this specifically. > > --- > > I am happy to continue the discussion of Axiom, Aldor and FriCAS but I > am not sure if this discussion is entirely appropriate the the > sage-devel list. Is there a Julia/Nemo list where this should be > continued? > > Regards, > Bill Page. > > > On 1 October 2015 at 10:56, Bill Hart <goodwillh...@googlemail.com> wrote: > > > > On Thursday, 1 October 2015 16:35:20 UTC+2, Bill Page wrote: > > <SNIP> > >> > >> Unfortunately while I am very much in favor of the category/domain > >> approach of Axiom and related systems, I find the Sage implementation > >> of this idea almost entirely indigestible. Perhaps this is not the > >> case for a sufficiently large number of potential Sage developers. > > > > Bill, I would be very interested if you could elaborate on this point in > > more detail (assume I don't know anything about Aldor/Spad because it is > so > > long ago that I read the manual for Aldor that I really have forgotten > how > > this works over there). > > > > I'm currently implementing a system in Julia which really follows the > design > > of Sage's parent/element setup quite closely (we tried another approach, > but > > it failed). Since the project I'm working on is yet very young, I'd like > to > > understand what is unpalatable about the Sage approach, since I'm > > essentially using it. > > > -- 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/d/optout.
