Wikidata:Requests for comment/DBpedia import process
An editor has requested the community to provide input on "DBpedia import process" via the Requests for comment (RFC) process. This is the discussion page regarding the issue.
If you have an opinion regarding this issue, feel free to comment below. Thank you! |
THIS RFC IS CLOSED. Please do NOT vote nor add comments.
- The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
- Stale. John F. Lewis (talk) 15:37, 14 May 2014 (UTC)[reply]
I have been to the DBpedia conference in Amsterdam. One frequent topic was how DBpedia and Wikidata could cooperate / coexist and how DBpedia can use its processes of its own update for Wikidata as well. What we came up with is more or less along these lines:
- The first step is data collection
- The second step is match the results to Wikidata and DBpedia.
- produce a lists with differences where Wikidata and the new data differ.
- update the new information to Wikidata / DBpedia simultaneously
- Ask the communities to reconcile the differences between the source Wikipedia and Wikidata
The other thing that is really relevant is that many people at the conferences represented GLAM partners of our movement. There were people from Europeana, the Dutch National Library among others. They are quite interested to run pilot projects on subjects like "Dutch authors" and "Dutch monuments". These would then be run along the lines as described above.
In one meeting I mentioned the Commons/Wikidata integration. I explained why the use of the "Creator" and "Institution" templates are a good step in preparation of this integration. I urged them to keep as much meta data as possible after an initial import so that we may make use of it when the Commons/Wikidata integration is a reality.
For DBpedia the way we have our "model" is/will be very much a challenge on so many levels. They have a big challenge to continuously keep a mapping to what our model happens to be. We do have qualifiers and processes to add but particularly remove properties. They already have a Wikidata project for Wikidata so they are aware of many of the issues.
Thanks, GerardM (talk) 14:12, 5 February 2014 (UTC)[reply]
- I did an analysis of the DBpedia ontology and posted it at Wikidata:Requests for comment/Can we reuse anything from DBpedia?. Their ontology is very different from that of Wikidata and I didn't see any easy way to import from there. No one posted any response to my comment however. Maybe you will get more response. Filceolaire (talk) 21:56, 7 February 2014 (UTC)[reply]
- The point is not that there are issues using their data, it is that they are going to have a process whereby they update our data using out a mapping that fits us from the data they harvested using their processes. It is in effect nothing we do not do ourselves. GerardM (talk) 12:20, 10 February 2014 (UTC)[reply]
-
- Filceolaire Gerard is right, the import process is just a question of mapping, nothing not doable with a little code. The mapping is challenging for the Dbpedia guys because the Wikidata ontology is more a matter of nonformal rules and convention than something that has clear formalized patterns, which means the automation of that process is not that easy. In clear, one bot to code specifically for each import case. TomT0m (talk) 17:49, 10 February 2014 (UTC)[reply]