GRAIL—A Generalized Representation and Aggregation of Information Layers

This paper was given at LAW XVI this morning, and I think many here will be interested. Presented here is another general-purpose data format for linguistic data, with particular focus on multimodality and rich multilayer annotation. The related work sections (2–4) are great, so that at the very least is worth reading for anyone interested in linguistic data formats. I also find the design discussion very interesting.

1 Like

I was at LREC 2014 when they started the LRE map effort. I had a talk in a workshop about clld and linked data, and had a hard time shoe-horning these language resources into the metadata scheme for LRE map (and in fact can’t find any of “my” resources easily anymore). WALS is in there a couple of times - highlighting that it’s used regularly in LREC papers.

So it might work as a metadata repository making explicit which data was used in which paper. But it doesn’t work as archive of language data. E.g. it doesn’t make sure stuff is available (my first search for a lexicon has a dead link LRE Map and no data on LRE map) and I also wouldn’t want WALS to be cited as LR deposited on LRE map by someone who used WALS data for a LREC paper.

Oh, and I never really understood the ISLRN thing.