Wikidata:Property proposal/Pressemappe 20. Jahrhundert IFIS

PM20 folder ID edit

Originally proposed at Wikidata:Property proposal/Authority control

   Done: PM20 folder ID (P4293) (Talk and documentation)
Descriptionidentifier for a folder in the Pressemappe 20. Jahrhundert of the German Pressearchiv 20. Jahrhundert
Represents20th Century Press Archives (Q36948990)
Data typeExternal identifier
Domaintopics, events, persons, companies and goods
Allowed values(pe|co)\/\d{6}|(sh|wa)\/\d{6},\d{6}
ExampleWerner Heisenberg (Q40904) → pe/007478 (Heisenberg, Werner)

BBC (Q9531) → co/002772 (BBC British Broadcasting Corporation (London))
petroleum in the United States (Q7178964) → wa/142108,141653 (Benzin : USA)

Baghdad Railway (Q639128) → sh/141113,161612 (Irak : Bagdadbahn)
Sourcehttp://webopac0.hwwa.de/PresseMappe20/
External linksUse in sister projects: [ar][de][en][es][fr][he][it][ja][ko][nl][pl][pt][ru][sv][vi][zh][commons][species][wd][en.wikt][fr.wikt].
Formatter URLhttp://purl.org/pressemappe20/folder/$1
Robot and gadget jobsMapping from GND ID (P227) will be possible for persons and corporations when http://purl.org/pressemappe20/beaconlist/pe and http://purl.org/pressemappe20/beaconlist/co is supplemented with the folder ID.
Motivation

Extensive archive of (German-language) press articles mostly from the 20th century (the offline version of the archive contains 20 millions of them, though most are not yet digitized). In dewiki, only a few entries are linked via de:Vorlage:Pressemappe, but the general usefulness of the archive has been discussed, so it sure would be a nice addition to Wikidata, too. Having the specific IDs (IFIS-Identnummern) is not strictly required, as a mapping for most (if not all?) entries to the GND exists (see above, robots jobs section; this is also the way they are linked in the said template), but obviouly the curators of the archive don't want to rely solely on that themselves. --YMS (talk) 11:33, 29 August 2017 (UTC)[reply]

@YMS: Are these IDs always six digits long? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 06:38, 5 September 2017 (UTC)[reply]
[1] only explains that they are unique and persistent. Actually, the example given on this page has only two digits (10), but if you click that link, it will be displayed on the page as 000010 nonetheless. That was true for all examples that I checked, but I can't guarantee that it's always true. And when checking this, I saw that th entries in the "Sacharchiv" do have two different IFIS umbers (e.g. "IFIS-Geo: 141595; IFIS-Sach: 145941"), and for none of them there's a link offered. --YMS (talk) 07:00, 5 September 2017 (UTC)[reply]
Discussion

Just a quick heads-up: I've put the proposal on hold, since my colleague is currently re-working the URLs. They will be based on Persistent Uniform Resource Locator (Q195305) and are intended to address a single folder of the person, corporation, wares or topic archives. (Later on, persistent URLs for single documents within the folders or for lists of folders may follow, which would probably require separate properties.) The implementation will take several more weeks. Jneubert (talk) 15:04, 13 September 2017 (UTC)[reply]

  • @YMS, ArthurPSmith, Pigsonthewing: My colleague has notified me that the new persistent identifiers for folders are in effect now. As documented here, they are based on purl.org. They consist of a two character prefix ("pe" for persons, "co" for corporations, "wa" for wares and "sh" for general subjects and events), a slash and either one or two six-digit ID number(s) from the IFIS system (since "wa" and "sh" folders are organized by subject + geography, for these the second ID is mandatory.) The identifiers can be obtained within the application from the icon named "Mappen-Zitier-Link".
I have adapted the proposal accordingly and added examples for the different cases. Furthermore, I suggest to change the name of the property to "PM20 folder ID", which leaves room for future properties which may address single documents or lists of folders. Jneubert (talk) 11:30, 5 October 2017 (UTC)[reply]