On Tuesday 8 October 2024 at 13:20:54 UTC-7 marc....@gmail.com wrote: On Tuesday, October 8, 2024 at 1:23:55 PM UTC-6 Nils Bruin wrote:
- the examples we have of bits of software developed as part of sage that ended up as library components of other projects are peripheral, interfacing parts that were spun off into independent libraries. - we don't have examples of core functionality of sage getting adopted as a component of another project The most obvious reason for this is that it too hard for a project to adopt core functionality of Sagewith the current organization of Sage. Making that easier was one of the main motivations for the modularization project. So this is circular reasoning. Indeed, it being hard to use the sagemath core in other projects would be an obstacle to doing so. However, it is not obvious to me that this is the only obstruction. It could be that there is simply no demand, regardless of how easy it is. And, as pointed out, licensing could be another obstruction. So it seems to me that the opportunity of benefit to sage from use in other projects is an unknown quantity at the moment. On the other hand, it seems there are other ways in which making sage more modular may bring benefits. As you're pointing out, sage still fits within 10GB in source, so it looks like sagemath could just be one pypi package. And yet, a debian packager is enthusiastic about being able to do it in smaller bits. Personally, I've seen sage develop as this monolithic, tightly interlinked library, where even the order of imports is delicate because of circularities. It seems to me that splitting that up comes at a considerable cost to actually perform. Particularly if, in addition, we get extra load imposed because all doctests now have to declare their prerequisites. To get buy-in from a lot of other developers, there will need to be a clear and obvious benefit to implementing this modularization. I think it will be much easier to get that buy-in by exhibiting some sure or nearly sure pay-off from the investment. The hope for a post-fact pay-off in the form of code reuse will probably be a bit too nebulous for many developers to put up with the extra costs. -- 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/463f8081-b2a0-4160-a58f-85324cce4f30n%40googlegroups.com.
