geneontology / go-ontology

Source ontology files for the Gene Ontology
http://geneontology.org/page/download-ontology
Creative Commons Attribution 4.0 International
220 stars 40 forks source link

(suggestion) Wikipedia is which Wikipedia? #18552

Open rwst opened 4 years ago

rwst commented 4 years ago

GO refers to pages in the English language Wikipedia simply as Wikipedia. Although not very probable the German language one could be given as reference too. So I'd propose to use the official Wikimedia database names which would be enwiki: and dewiki:, respectively. You can find a list of all Wikimedia database names at: https://www.wikidata.org/w/api.php?action=sitematrix

ukemi commented 4 years ago

@hdrabkin, if we decide to go this route, it might need to be reassigned to someone who can do it with a script.

hdrabkin commented 4 years ago

this could almost be done with a simple find/replace maybe using a text editor, then loading back into protege to make sure everything's ok, like I did with the plant-zymes. However, this also means we would need to have two dbx for any one wiki. I don't see the point. Frankly, PMIDs go the ncbi; it's in English.

srengel commented 4 years ago

would be good to not use wikipedia at all for GO.

rwst commented 4 years ago

would be good to not use wikipedia at all for GO.

1092 references in GO: 750 in xref, 264 in def, 78 in synonym. Wikipedia has references for what they write. It's no longer the early days of schoolkid vandalism (2005?) because authors are observed for some months before they can write now.

rwst commented 4 years ago

This ticket is only about renaming "Wikipedia:" to the officially correct "enwiki:" like GO uses other official database tags.

ValWood commented 4 years ago

Although Wikipedia is fab and highly accurate agree with @srengel we probably shouldn't use as a citation for defs because it isn't static. It will be updated as scientific opinion changes. We should use the sources that Wikipedia cites. That said, it isn't a good use of GO time to replace existing ones!

hdrabkin commented 4 years ago

Definitely agree @ValWood ;

cmungall commented 4 years ago

We discussed on the editors call

To echo @rwst's comment:

This ticket is only about renaming "Wikipedia:" to the officially correct "enwiki:" like GO uses other official database tags.

We agree - there is a lot to say about existing definition provenance in GO (not just wikipedia, but all the URLs that no longer resolve), but let's open a new ticket/project for this.

We will change definition xrefs to enwiki in the ontology. We will also update db-xrefs.yaml to add a new entry for enwiki, such that enwiki will resolve to the English language wikipedia

balhoff commented 4 years ago

@cmungall I thought the conclusion was to change all Wikipedia xrefs to enwiki, not just definition xrefs. Do you agree?

cmungall commented 4 years ago

@balhoff - I think there are arguments to be made either way. The xref refers to a concept that should presumably be equivalent across languages, the definition is sourced from text that has a language-specific expression

But yes I think it's probably simplest to change all xrefs. We may want to replace the class xrefs (concept level) with references to wikidata IDs in the future.

rwst commented 4 years ago

If you want to link to the language-independent concept behind a Wikipedia article, link to its associated Wikidata item. This also contains the full set of links to Wikipedias where that concept has an article. Incidentally, Wikidata contains all GO items so you have an easy check that you have linked to the right item. Example:

id: GO:0001534
name: radial spoke
xref: Wikipedia:Radial_spoke

would be https://www.wikidata.org/wiki/Q7280278 with the canonical identifier wikidata:Q7280278.

Wikidata is a free and open knowledge base that can be read and edited by both humans and machines.

rwst commented 4 years ago

I can provide a diff replacing Wikipedia with Wikidata identifiers if this is wanted.