Wikidata:Requests for permissions/Bot/Research Bot
(Redirected from Wikidata:Requests for permissions/Bot/Research Bot 1)
- 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.
Approved--Ymblanter (talk) 18:06, 24 October 2016 (UTC)
Research Bot (talk • contribs • new items • new lexemes • SUL • Block log • User rights log • User rights • xtools)
Operator: Daniel Mietchen (talk • contribs • logs)
Task/s: Widar edits
Code: N.A. (see below)
Function details: I have been using QuickStatements (Q20084080) to work on large numbers of items and properties around scholarly publications, taxa, diseases, compounds and other research-related matters. I intend to continue to do so, and inspired by User:Danmichaelo's Bot 6, I am applying for the bot flag for this account in order to avoid flooding Recent Changes. --Daniel Mietchen (talk) 16:59, 17 October 2016 (UTC)
Support --Succu (talk) 17:23, 17 October 2016 (UTC)- Support --Pasleim (talk) 17:46, 17 October 2016 (UTC)
- @Daniel Mietchen: this is a good thing, but it would be better if you could add references for the statements on these items - obviously in some sense the DOI is a reference, but do you have a source database that you are pulling the information from (I'm looking at recent scholarly article entries). You don't need to add a source for the DOI or the instance of (P31) statement but for others (particularly the author names) it would definitely be helpful. Also is there some reasonable default "description" you could add (at least "scientific article, but maybe it should include a rough citation")? Otherwise when one of these comes up in the search box it might be unclear what it is. Also if you happen to add two articles with the same title (this does happen rarely) they should have differing descriptions. ArthurPSmith (talk) 19:51, 19 October 2016 (UTC)
- @ArthurPSmith: Thanks for your comments — I think they are of a kind that would be most welcome over at WikiProject Source MetaData. The items I am starting for scholarly articles normally have at least one of DOI (P356) or PubMed publication ID (P698) or PMC publication ID (P932); for biomedical ones typically all three. Given that the respective values for those properties lead directly (by way of formatter URL (P1630)) or indirectly (via source website for the property (P1896)) to a place where the information can be verified (example), I am not sure adding additional references like stated in (P248): PubMed Central (Q229883) (perhaps with some reference URL (P854) statement) is helping much in terms of verifiability. The tool that I am using for this is Magnus Manske's Source, M.D., which turns the author names obtained from the respective databases into author name string (P2093) statements. Turning these into author (P50) statements (e.g. by way of the dedicated Wikidata Game) is where the real value of Wikidata comes in (e.g. allowing to generate a list of authors of scientific articles who received a Nobel prize), but these P2093-to-P50 conversions are hard to reference, and I expect the approaches around that to evolve over time. For instance, I like the way User:Sarilho1 has modeled this here but haven't seen it used more widely, nor discussed anywhere, nor implemented in any tools I could use. As for the description of items about scholarly articles, I am aware of the problem and have been using SPARQL queries (example) to identify such cases and then add descriptions (example), often by way of Quick Statements, i.e. in the way that I intend to use the bot account. --Daniel Mietchen (talk) 21:20, 19 October 2016 (UTC)
- Ok, sounds like there are some complications and discussions already about this over in the wikiproject. I'll add a discussion item there but this isn't a reason to delay giving the robot flag here, what it's doing is good work and if it's decided we want more in the way of references that shouldn't be hard to add later. ArthurPSmith (talk) 15:02, 20 October 2016 (UTC)
- @ArthurPSmith: Thanks for your comments — I think they are of a kind that would be most welcome over at WikiProject Source MetaData. The items I am starting for scholarly articles normally have at least one of DOI (P356) or PubMed publication ID (P698) or PMC publication ID (P932); for biomedical ones typically all three. Given that the respective values for those properties lead directly (by way of formatter URL (P1630)) or indirectly (via source website for the property (P1896)) to a place where the information can be verified (example), I am not sure adding additional references like stated in (P248): PubMed Central (Q229883) (perhaps with some reference URL (P854) statement) is helping much in terms of verifiability. The tool that I am using for this is Magnus Manske's Source, M.D., which turns the author names obtained from the respective databases into author name string (P2093) statements. Turning these into author (P50) statements (e.g. by way of the dedicated Wikidata Game) is where the real value of Wikidata comes in (e.g. allowing to generate a list of authors of scientific articles who received a Nobel prize), but these P2093-to-P50 conversions are hard to reference, and I expect the approaches around that to evolve over time. For instance, I like the way User:Sarilho1 has modeled this here but haven't seen it used more widely, nor discussed anywhere, nor implemented in any tools I could use. As for the description of items about scholarly articles, I am aware of the problem and have been using SPARQL queries (example) to identify such cases and then add descriptions (example), often by way of Quick Statements, i.e. in the way that I intend to use the bot account. --Daniel Mietchen (talk) 21:20, 19 October 2016 (UTC)
