On Wed, Oct 2, 2024 at 6:54 AM 'Martin R' via sage-devel <sage-devel@googlegroups.com> wrote: > > So, it turns out that > > * a core distribution with absolutely minimal dependencies and only > > dependencies which have proved stable on all supported platforms > > > > This is exactly sagemath-categories. It has absolutely minimal dependencies > > is not what I tried to say. What I tried to say is that the core > distribution should have minimal dependencies, but contain as much of sage as > possible. > > I don't see any advantage of restricting to a smaller subset. What do you > have in mind people would do with just the stuff in `categories`?
With sagemath-categories as it is now, one can play a bit with polynomials, and (I guess) Sage's Sets. As one cannot do anything meaningful with sage.categories alone, as it's just Sage's "abstract nonsense", without any implementations of objects in these categories, sagemath-categories distro contains some more or less random, concrete, implementations of maths objects, e.g. Sage's polynomial rings. (And so a substantial part of sage.rings, I presume) So how is this absolute minimum? What if the user doesn't need polynomials? It must be, by the current design of distributions, imported in any Sage-dependent piece of code - by virtue of being in almost every sagemath-* distribution. By the way, Sage's categories is a bit of misnaming, I think --- but that's another story. It's not really an implementation of a bit of category theory in the mathematical sense, it's actually just types, in the CS sense, as we don't have there (not in the generality of category theory for sure) anything like functors, natural transformations, Yoneda lemma, etc. That's Sage's type system that's there in sagemath-categories, more or less. And so let's not add there semi-random pieces, unless it's renamed. Dima > > Martin > > On Wednesday 2 October 2024 at 02:25:38 UTC+2 Matthias Koeppe wrote: > > On Sunday, September 29, 2024 at 1:30:38 AM UTC-7 axio...@yahoo.de wrote: > > > I think that `sagemath--minimal-dependencies` would be clear, wouldn't it? > > > Again fails the "minimal for what" test; and fails to distinguish > sagemath-categories from sagemath-objects (which has the same dependencies). > > > I have now two more questions: > > * why are some classes without any dependencies missing from what's currently > named `sagemath-categories`? I don't know how to do a systematic search, so > I only checked the ones I know well, and `GrowthDiagram` is not there, but is > pure python. > > > "Some classes are missing" is really far from a description of > sagemath-categories. It's really much, much smaller than that. > > GrowthDiagram, like most of sage.combinat, will be shipped in > sagemath-combinat (which sits on top of sagemath-categories in the > dependencies graph). > > > * wouldn't it be better to tag dependencies, rather than having to explicitly > put something into what's currently `sagemath-categories`? I imagine that it > will be a constant source of confusion whether to put some new functionality > into all.py or all__sagemath_categories.py > > > There's no guesswork involved, and there is no potential for confusion > because the source files from which stuff is imported are tagged by "# > sage_setup: distribution" lines. It just has to be kept in sync; and the > automated tests ensure that. > > (I'll also note that the "all.py" files are in very poor shape in Sage; in > the past, too little thought has been put into what belongs into these files. > See Meta-ticket: deprecate useless or misleading functions in the global > namespace, https://github.com/sagemath/sage/issues/25383) > > -- > 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/5123f648-b830-4010-b016-5c1afd687b25n%40googlegroups.com. -- 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/CAAWYfq2Z%2BSm8mt%2BmU_h4pjN1gwwBbcai_aH4K4r-t7-pn62awg%40mail.gmail.com.
