Wikidata:Property proposal/Uniform Resource Identifier Scheme
Jump to navigation
Jump to search
uniform resource identifier scheme[edit]
Originally proposed at Wikidata:Property proposal/Natural science
| Description | IANA-assigned uniform resource identifier scheme which identifies items within this class or set |
|---|---|
| Represents | Uniform Resource Identifier scheme (Q37071) |
| Data type | External identifier |
| Domain | set (Q36161) (any class which contains items which may be referenced by a URI scheme), and as a particular example: network service (Q1640628) |
| Allowed values | regex: [a-z][a-z0-9+\-.]* |
| Example | filename (Q1144928) → file, website (Q35127) → https, email address (Q1273217) → mailto, telephone number (Q214995) → tel, IRC server instance (Q18511644) → irc, |
| Source | Uniform Resource Identifier (URI) Schemes |
- Motivation
Whilst there aren't many URI schemes currently registered, there are numerous other scheme identifiers used unofficially by software and common practice. URIs are extensively used (think http://, https://, mailto://, ssh:// and others) and it is therefore useful to record the formal structure of such schemes in Wikidata. Pixeldomain (talk) 01:29, 5 April 2017 (UTC)
- Discussion
- In principle this sounds like a good idea. In practice I'm not sure what most of the wikidata items would be for those schemes other than the most heavily used ones - for example for Andrew File System (Q504669), we have 'afs', but does a file system really define a protocol? And some of the others seem even less clear (what item would you link to "mailto"?). On the other hand, would it be helpful to apply the same property to URN namespaces, rather than add yet another property to cover those also? See this list (the identifier would then be urn:issn, for example for the issn namespace). ArthurPSmith (talk) 19:09, 5 April 2017 (UTC)
Comment @ArthurPSmith: My intention was that this property would be applied to classes of items which a URI is intended to reference, so for the "mailto" URI, property/value pair would be applied to email address (Q1273217) as this is the class of item intended to be referenced by a "mailto" URI. I see the confusion with my original proposal, as I had a bad example for the "http" scheme. This is now fixed. I can see how a "URI" property proposal (not "URI scheme") could also be created, with existing properties official website (P856) and email address (P968) (amongst others) becoming children of the new "URI" property. This "URI" property could be applied to any item in Wikidata for which a URI exists that references it. An existing example of a related property which already uses this approach is ITU/ISO/IEC object ID (P3743). Pixeldomain (talk) 02:34, 12 April 2017 (UTC)
- Well, ok, that makes sense for mailto and http/https. But then what about ssh, what are you representing there? Remote shell operations? ftp? afs as I mentioned - files? ipp - representing a printer? ldap representing a directory?? I think it would be helpful to work out the mapping you are thinking of here for a dozen or so examples to be sure this makes sense. ArthurPSmith (talk) 12:30, 12 April 2017 (UTC)
- Comment looks good but needs further elaboration (examples, constraints...) as already commented by ArthurPSmith -- JakobVoss (talk) 20:11, 11 November 2017 (UTC)
Support
--- Jura 19:04, 9 December 2017 (UTC)- Support --Okkn (talk) 20:46, 30 December 2017 (UTC)
- Comment @ArthurPSmith: @Jura1: I have added a few more examples, clarified the description and provided some clarification on the domain. Pixeldomain (talk) 01:08, 10 January 2018 (UTC)
- It sounds like you have a few more cases that this makes sense for but I'm still a little doubtful regarding filename (Q1144928) here - why filename and not file or file system or something more service-related? And none of these relationships seem to be described directly on the IANA list you reference - what fraction of the 200 or so URI prefixes could actually be mapped in this way? Do the browser-extension URI's make any sense to support with this? @JakobVoss: what do you think now? ArthurPSmith (talk) 14:17, 10 January 2018 (UTC)
- @Pixeldomain:
Done I don't think we need to sort out beforehand every item that gets this property.
--- Jura 08:09, 12 January 2018 (UTC)
