Closed oruebel closed 5 months ago
I like Linked Data Interface, but I worry that it will be overly broad and connected to JSON-LD. We are really interfacing between chunked array storage. How about Linked N-Dimensional Array Interface (LINDI)?
Some other options that didn't quite make the cut:
Or zarrbie :)
Or we could call it LINDI and have it stand for multiple things, depending on the context.
Distributed Archives for Neurophysiology Data Integration = DANDI Linked Neurophysiology Data Interface = LINDI
and then when it's not neurophysiology
Linked Data Interface
So far, my personal favorites are:
All right, let's go with Linked Data Interface (LINDI). It's growing on me.
The README currently defines the name as Linked Neurodata Interface (LINDI). I'm wondering how specific the library actually will be to neurodata. It seems this library may be much more broad in nature, in that it only cares about the data format (HDF5, Zarr, etc.) but it may not really care about the actual schema. Dealing with the schema and application-specific functions, seem like they will likely remain with the domain libraries, e.g,. PyNWB, HDMF etc.. With that in mind (and keeping with the current acronym), I'm wondering whether it should be just Linked Data Interface (LINDI).