They could very well all end up in one place, but as long as they are maintained and work together seems to be more important to me.
1 Like
BTW. I think there are a lot of cool things each has to offer and I’d love for there to continue to be an open dialog about what people have created and are willing to work together towards. I really only created AbstractIndices as a proof a concept for a PR to IndexedDims because it’s hard to explain how a lot of it just magically works with base until you try it. My break down of what is most beneficial from each package is this:
- NamedDims: simple low allocation named dimensional manipulation. It’s clean, fast, and if you follow the conversations on github they always work towards the most intuitive solutions for users.
- IndexedDims: Not really working yet but I hope it just focuses on a very generic method of tying AbstractVectors to the base indexing system. People will always want to create new frames of indexing, so the less we have to plan ahead and create those now, the better.
- DimensionalData: really, really, cool ideas on how to use semantic information to interface with an index and dimension. The ability to refer to something programmatically as “near” or “between” is something I’ve yet to find well defined in Julia. It has a lot of the tools necessary to tie the first two packages together to easily create domain specific interfaces.
7 Likes