Wikidata:Property proposal/ politician identifier

From Wikidata
Jump to navigation Jump to search politician identifier[edit]

Return to Wikidata:Property proposal/Organization

   Under discussion (the official site of the Danish parliament) has an identifier string in the URL's of biographies of politicians. Politician's items can use this property and e.g. Wikipedia can automatically link the official biography on
Data typeString
Domainmember of the Folketing (Q12311817)
Allowed values[a-z-]+ (note that the a-z should include Danish and Færøer characters)
Example 1Carolina Magdalene Maier (Q20197678)carolina-magdalene-maier
Example 2Per Stig Møller (Q697184)per-stig-møller
Example 3Sjúrður Skaale (Q845278)sjúrður-skaale
Planned useautomatically generate links from English and Danish Wikipedia articles of politicians
Number of IDs in sourceat least 179 current members of Folketing, probably double for all former members of the last decade
Expected completenessalways incomplete (Q21873886)
Robot and gadget jobsYes, should validate that the implied URL is not a 404.
See alsoProperty:P6849


Currently, Danish and English (and presumably other) Wikipedias manually link the politicians official biography. All recent politicians have profile pages like this, the Parliament management writes the biography. By having a property, it would be easier to change the URL prefix, it would be easier to update values across Wikipedias, the value could be automatically used from e.g. da:Template:Infoboks MF. That infobox already shows official URL's, but they are freeform and some politicians have the official bio in there, some do not (for arbitrary reasons). By using this property, information would be organized in a stricter schema. Note that politicians that retired long ago do not have current biographies (their bio's have different URL schemes). This proposal does not concern those politicians.

The different classes of biographies on are (as far as I can see):

newish (note that she also has a bio with the current schema . Her item would have this property with value "merete-dea-larsen")
old (note that she does not have a bio with the current schema or the other old schema. Her item would not use this property, manual linking would continue)
new (note that she ONLY has a biography with the newest schema. This search with the foreign URL scheme yields no results: [1]. Her item would have this property with value "aki-matilda-hoeegh-dam")

There are also PDF versions, sometimes in English, but I don't think we should link those, they offer no inherent advantage over HTML.

If any biography page on loads with no text, set the language to Danish by prepending "/da" to the path portion of the URL.

Note that the part before the last / in the URL is just the first letter of the name. For example, Carolina has a "c" in the URL as seen above. This character can be extracted from the property value by Mediawiki string manipulation functions using Scribunto.

I would set the "formatter URL" to{{Module:String|sub|s=$1|j=2}}/$1 but the proposal template doesn't accept my syntax with template invocations and nowiki tags.

Ysangkok (talk) 18:32, 6 July 2019 (UTC)


Is there any way to combine the two versions in the same property? The identifiers can be extracted from each other, but the rules for that cannot be expressed in the formatters as far as I know. --Dipsacus fullonum (talk) 21:26, 7 July 2019 (UTC)
Oh, didn't realize it was so simple. I propose we just only save the part after the slash, e.g. "carolina-magdalene-maier". Templates can use Lua to take the first character, and generate the Danish bio link programmatically. It can easily be done using e.g. w:Module:String. I am not sure why you are doubting this, since you know about how templates work. As previously mentioned, it would be Wikipedia templates actually generating the links, Wikidata just needs the property so that we'd be able handle the eventual case of a politician where the ID on doesn't match the name we store. Actually, we'd probably be able to generate 99% of the links by just taking all names, concatenate them with "-" in between, and prepend the prefix (with the Danish link having the extra duplicated first character). But I would prefer to have this property, since it would be fragile without it. --Ysangkok (talk) 17:48, 8 July 2019 (UTC)