Tune up README.md

[yarikoptic]

I have provided initial summary in the README now, so we have our (sorry if there is some personal bias in what I have written so far) ideas. Please contribute! Now that there is some "meat" in that file, PRs might be sensible and provide a forum for discussion

[magland]

Looks great @yarikoptic. Are you thinking of PRs representing edits on the README.md?

[satra]

for the datasets, could we just make them part of dandi-datasets? that way we address both connecting to dandi to evaluate the spikes and showing how other datasets can be added to dandi.

[yarikoptic]

Looks great @yarikoptic. Are you thinking of PRs representing edits on the README.md?

I don't mind direct pushes, but with PRs we could discuss any changes if submitter feels that there could be something to discuss ;-)

[yarikoptic]

for the datasets, could we just make them part of dandi-datasets? that way we address both connecting to dandi to evaluate the spikes and showing how other datasets can be added to dandi.

the beauty of datalad -- we could have them everywhere ;)
or more seriously, I envision datasets originate somewhere, and where I do not really care (could be dandi-datasets, or elsewhere where it is reliable) but we would collect here under https://github.com/spikodrome/datasets the superdataset of datasets of interest (raw + curated data; or just alternatively curated data). Most likely greedy me would also host them (all, with or without data where unfeasible; that is where data could better also reside on DANDI) under ///spikodrome/datasets of http://datasets.datalad.org/ for extra redundancy.

[magland]

I like the idea making use of datalad to host the same datasets in multiple places, esp. during development stage.