Re: webid predicate

Nb: https://www.wired.com/story/google-wants-to-kill-the-url/

As TimBL has noted, URIs (including DIDs) are essential.

Timo


On Wed., 19 Sep. 2018, 11:28 pm Melvin Carvalho, <melvincarvalho@gmail.com>
wrote:

>
>
> On Wed, 19 Sep 2018 at 15:22, Story H.J. <H.J.Story@soton.ac.uk> wrote:
>
>>
>>
>> > On 19 Sep 2018, at 09:34, Melvin Carvalho <melvincarvalho@gmail.com>
>> wrote:
>> >
>> >
>> >
>> > On Wed, 19 Sep 2018 at 09:25, Story H.J. <H.J.Story@soton.ac.uk> wrote:
>> >
>> >
>> >> On 19 Sep 2018, at 00:08, Melvin Carvalho <melvincarvalho@gmail.com>
>> wrote:
>> >>
>> >> Hi All
>> >>
>> >> This was raised a few years back, but not resolved.
>> >>
>> >> We dont have a predicate for webid itself.
>> >>
>> >> It would be useful in heterogeneous systems to write
>> >>
>> >> []  :webid <myPreferredURI> .
>> >
>> > by using <myPrefferedURI>  you are speaking of the thing, not the URI.
>> > <-> is a function from String to URIRef, written mathematically as:
>> >
>> >     <->: String → URIRef
>> >
>> > This means it can be substituted with co-referring terms. So that's not
>> what you want.
>> > You'd need
>> >
>> > person :webid "http://bblfish.net/people/henry/card#me"^^xsd:anyURI .
>> > to refer to the URI as a URI.
>> >
>> > What would be the domain and range of that relation?
>> >
>> > So, the use case is geared towards adding a URI to an existing system.
>> Say having "webid" : foo in your git config (which hopefully could be
>> integrated into git fully).  Or to put it in matodon, anything that allows
>> you to tie key value pairs to an existing profile.  The benefit is to pull
>> web 2.0 systems into linked data using follow your nose.
>> >
>> > As such I think the domain has to be anyURI.
>>
>> You can follow your nose with xsd:anyURI. It requires a special step -
>> easy to program - but it won't be merged by
>> reasoners, for which we may as well build now.
>>
>> >
>> > As for the range, I think it could be foaf Agent, but also there's a
>> number of other agent type vocabs as of 2018 e.g. schema.org, as, ogp --
>> im sure a few more, and more will emerge.  So I think range ought to be
>> anyURI too?
>> >
>> >
>> >>
>> >> Let me give you a use case.  In th popular web 2.0 federated
>> microblogging service, mastodon, you are allowed to specify key value pairs
>> to tie to your profile (nice!).  So what would be valuable would be to add
>> a webid, with the appropriate label so that it is possible to follow your
>> nose to the linked data universe.
>> >>
>> >> The question is where to put it.
>> >>
>> >> Strangely enough webid doesnt actually have an identity vocab.  Should
>> we start one perhaps?
>> >>
>> >> It could also go on the cert ontology, tho not quite the right place.
>> >>
>> >> It could also go in the solid vocab, which seems kind of OK.  For
>> various reasons it's better to have it on one place, I think.
>> >>
>> >> As a stop gap we could use the preferredURI predicate, tho unsure
>> that's ideal ..
>> >
>> > I forgot where that is defined? How is it defined there?
>> >
>> > https://lov.linkeddata.es/dataset/lov/terms?q=preferredURI
>>
>>
>> $ curl --silent  -H "Accept: text/n3"
>> http://www.w3.org/2000/10/swap/pim/contact | grep -A3 preferredURI
>> :preferredURI a rdf:Property;
>>     rdfs:label "preferred";
>>     rdfs:comment """A string which is the URI a person, organization,
>> etc, prefers that people use for them.""".
>>
>>
>> It has an informal definition which states that the domain is a
>> foaf:Agent and the range a String.
>> That's Tim Berners-Lee's ontology. Seems to me that one could just
>> formalize the definition there.
>>
>
> Here's a bit more background on that :
>
> https://www.w3.org/DesignIssues/QuotingURIs.html
>
> Good suggestion, we could put it in pim : contact -- but who owns that
> namespace.
>
> Another thought would be to have https://w3id.org/ with webid -- but that
> would be less consistent with what's existing in webid
>
> I like nathan's suggestion of creating a new vocab.  Would need to look
> into the logistics of publishing and maintenance tho.
>
> >
>> >
>> >>
>> >> Any thoughts?
>>
>>