Add section on tooling

[saulshanabrook]

I wrote up some of the motivation and details behind the record-api tooling.

Happy to expand or reorganize from any feedback.

I didn't add many details about how the array-api-comparison works. @kgryte feel free to lemme know what I should add there, or add that in a follow up.

[kgryte]

@saulshanabrook Nice draft! Left some comments (notably, need to turn on IDE spellcheck ;)).

I think a larger question concerns how much implementation detail we actually want to describe, given its likelihood to change and a desire to minimize spec "noise".

[saulshanabrook]

Thanks @kgryte!

Nice draft! Left some comments (notably, need to turn on IDE spellcheck ;)).

Yes, need to do this! Was just writing in my browser...

I think a larger question concerns how much implementation detail we actually want to describe, given its likelihood to change and a desire to minimize spec "noise".

I took out some of the details, but I do think this will still need to evolve a bit as our data process continues to evolve.

[jorisvandenbossche]

(sidenote: the aspects that are too detailed here, might be interesting to move to the python-record-api repo itself? Eg to its README or docs)

[rgommers]

Thanks @saulshanabrook, this looks like a good start to me.

Regarding structure, we have two tooling topics, so we need to decide whether to keep the structure of sections as it is now, or to split into the array-api-comparison and python-record-api parts right after the summary. I think I'd favour the latter.

There's probably a bit missing about how we use the data for decision making. That should come at the end I think, after it's clear to the reader what data is produced.

Also, can you wrap to 80 char lines? That will make it much easier to review.

[saulshanabrook]

Regarding structure, we have two tooling topics, so we need to decide whether to keep the structure of sections as it is now, or to split into the array-api-comparison and python-record-api parts right after the summary. I think I'd favour the latter.

That makes some sense. I haven't started this yet.

There's probably a bit missing about how we use the data for decision making. That should come at the end I think, after it's clear to the reader what data is produced.

I added a few sentences on this, but wasn't very specific.

Also, can you wrap to 80 char lines? That will make it much easier to review.

Sure, wrapped.

[kgryte]

@rgommers This should be ready for another round of reviews.

[saulshanabrook]

Looks good to me!

[rgommers]

LGTM too, just pushed a couple of tiny textual fixes. The story flows very well here, and the section ends up being nice and concise.