Open goodmami opened 5 years ago
I think the best way to do this would be to create a separate file for the proposed links, e.g., ili-proposed.ttl
and include the links there. I think this would avoid any confusion about which identifiers are confirmed and which are still in development
ili-proposed.ttl
, ili.ttl
, and ili-deprecated.ttl
(this issue)Specifying superseding concepts on deprecated ones is orthogonal to the issue of indicating status; either of the above would need some solution for this.
I am happy to split proposed ilis into a different file, but I don't see why we should split out deprecated ilis. We can assume that wordnets will not always change the moment ILIs change, so the they would always need to have access to the deprecate ilis, no?
In any case, I think we should mark when a new ili was proposed, and I am happy to also mark it as proposed.
I think defining our own labels (deprecated, proposed, active [= default or no tag]) is probably a good idea, I am not sure of the best way to do this in terms of RDF.
We can assume that wordnets will not always change the moment ILIs change, so the they would always need to have access to the deprecate ilis, no?
I don't see how splitting them into files mean the wordnets lose access to them. But if you want them all in a single file, one solution is that the exported TSV file for downstream applications combines all these files with a column indicating the status, so all ILIs are available to those applications.
In any case, I think we should mark when a new ili was proposed, and I am happy to also mark it as proposed.
Thinking about the when... We currently have dc:source
which refers to a particular wordnet version. If we need finer-grained times than that we could add some datetime instance to each entry, but I think that might be overkill. Let's just say dc:source
is the wordnet version that proposed the ILIs?
I think defining our own labels (deprecated, proposed, active [= default or no tag]) is probably a good idea, I am not sure of the best way to do this in terms of RDF.
Me neither. If we want a status property in RDF instead of split files, let's discuss it on #8.
When a wordnet is uploaded to OMW, new CILI concepts are granted a temporary "proposed" status ("proposed" is just illustrative, the terminology is not fixed; maybe "candidate"?). After a month, the concept may be accepted or deprecated if the wordnet is rejected (and in the future there may be a mechanism for deprecating individual ili concepts as well). If a file in this repository is generated while the concept has this status, the status should be encoded somewhere so it can indicate to users that the concept is not yet "official", and they can then decide whether or not to use the concept.