I like notion of writing down “user paths”, that’s interesting.
I was thinking of a couple that I have used myself:
find an interesting open resource (CLDF, Wikipedia/-media/-etc, a repo, a reference in a paper…) → figure out something to do with it
Heh. I mean, when I find something that’s open that’s often my reaction; I almost never assume that there will be open resources for a given language. Maybe with time expectations will change.
Another relevant thing is that for people who specialize in poorly documented languages, they mostly know what exists and will have a good sense of whether there’s going to be anything digital at all. I think even there, though, sometimes such people will look for stuff (along the paths listed above) for materials on related languages. If they can’t find anything closely related, they go up a node and look for the next family up (if there is one).
(Random thought: I wonder if would be useful if there were some kind of recognizable bibliographic reference format that would be understood to link only to data itself, as opposed to (or better, alongside) other kinds of resources. )
From my POV time would be an essential dimension of this network - so maybe adding some sort of indication about “latest activity”, presumably derived from “last updated” timestamps in the data would be good (and even this may be inaccurate because sometimes servers just keep on running …).
That’s true… I haven’t played around with the time dimension. time is also really hard because some data providers fall off the rid and then come back. There is also size of the number of entries provided. My purpose here though is to look at technologist implementing. are there clld apps you’ve deployed which don’t act as data providers to to OLAC?
Yes, tons Most of the 83 repos in Cross-Linguistic Linked Data · GitHub correspond to (deployed) clld apps. All of these could act as OAI-PMH providers for OLAC, but most are not registered; mostly because I’m not sure how useful this would be.
)
Most of the 83 repos in