User friendly metadata identification

[saleiro]

It's very hard to communicate using hashes. This issue proposes to assign a user friendly identifier (can be a int just like model_id and model_group_id) to the experiment, subset and matrices metadata.

[jesteria]

I was kind of wondering about that myself. Hashes and UUIDs aren't crazy – we use them in Git, of course, and they work fine – but they are hard on human brains.

Is there any need for universal uniqueness, etc.?