Re: Profiles in Linked Data

I think there's the issue of the data format but also the problem that

<http://someorganization.org/namespace/something>

might represent someorganization.org's viewpoint about :something if you
are lucky.  On the other hand you might want to know what somebody else
thinks about that "thing"  -- i.e. you might want to follow a dbpedia
identifier to get schema.org information on it that somebodyelse.net has
compiled on the dbpedia universe.

I.e. fundamentally dereferencing has to be extended to support the idea of
"what does source X think about resource Y?"

There is also the need to recognize that dereferencing has created a lot of
confusion.  I suspect some people have been intimidated from using RDF
because they think that having names based on URLs means that they *have*
to publish everything on the web.

On Mon, May 11, 2015 at 3:14 PM, Kingsley Idehen <kidehen@openlinksw.com>
wrote:

> On 5/11/15 11:54 AM, Svensson, Lars wrote:
>
>> Kingsley,
>>
>> On Saturday, May 09, 2015 12:07 AM, Kingsley Idehen wrote:
>> [...]
>>
>>> So to repeat my question in another mail: I have an entity described by a
>>>> (generic) URI.
>>>>
>>> You have an entity identified by a IRI in RDF. If you are adhering to
>>> Linked Open
>>> Data principles, said IRI would take the form of an HTTP URI.
>>>
>>>
>>>    Then I have three groups of documents describing that entity, the
>>>> first uses
>>>> schema.org, the second group uses org ontology and the third uses foaf.
>>>>
>>> You have an entity identified by an HTTP URI. The dual nature of this
>>> kind of
>>> URI enables it function as a Name. The fundamental quality (attribute,
>>> property, feature) of a Name is that its interpretable to meaning ie., a
>>> Name
>>> also has a dual (denotation and connotation feature) which is what an
>>> HTTP URI
>>> is all about, the only different is that denotation->connotation (i.e.
>>> name
>>> interpretation) occurs in the hypermedia medium provided by an HTTP
>>> network
>>> (e.g. World Wide Web). Net effect, the HTTP URI resolves to and document
>>> at a
>>> location on the Web (i.e, a document at a location, which is the URL
>>> aspect of
>>> this duality).
>>>
>> OK. I have an http URI that denotes an entity. Depending on the server
>> configuration and what accept-headers I provide, the http dereferencing
>> function returns a document at a location.
>>
>>  All documents are available as RDF/XML, Turtle and xhtml+RDFa. How does a
>>>> client that knows only the generic URI for the resource tell the server
>>>> that it
>>>> prefers foaf in turtle and what does the server answer?
>>>>
>>> It can do stuff like this:
>>>
>>> curl -L -H "Accept:
>>> text/xml;q=0.3,text/html;q=1.0,text/turtle;q=0.5,*/*;q=0.3" -
>>> H "Negotiate: *" -I http://dbpedia.org/resource/Analytics
>>>
>> OK, I can see how setting the Accept-header negotiates the media type. If
>> I understand correctly, the Negotiate-header gives the server and
>> intermediate proxies a carte blanche to negotiate things any way they
>> prefer. I don't see any header that tells the server what
>> profile/shape/vocabulary the client prefers.
>>
>
> That's about a client negotiating different types of document content
> using a preference algorithm which in integral to Transparent Content
> Negotiation. It has nothing to do with a preferred vocabulary of terms
> e.g., dcterms vs schema.org in regards to terms used to describe
> something using RDF Language bases sentences/statements.
>
> If you want an RDF based entity description, where the terms used come
> from a specific vocabulary, that's where you could leverage a query
> language e.g., SPARQL. Of course, there are those that don't want to use
> SPARQL which could then lead to yet another kind of "profile" relation
> object, but ultimately such use will only be the equivalent of ignoring the
> existence of "multiplication" and "division" in regards to arithmetic
> operations.
>
> Conclusion: if folks want to build "profile" relations for selecting RDF
> content constructed using terms from a specific vocabulary, that's fine
> too, even though its  utility would simply boil down to navigating politics.
>
>>
>>  HTTP/1.1 303 See Other
>>> Date: Tue, 05 May 2015 16:01:06 GMT
>>> Content-Type: text/turtle; qs=0.35
>>> Content-Length: 0
>>> Connection: keep-alive
>>> Server: Virtuoso/07.20.3213 (Linux) i686-generic-linux-glibc212-64  VDB
>>> TCN: choice
>>> Vary: negotiate,accept
>>> Alternates: {"/data/Analytics.atom" 0.500000 {type
>>> application/atom+xml}},
>>> {"/data/Analytics.jrdf" 0.600000 {type application/rdf+json}},
>>> {"/data/Analytics.jsod" 0.500000 {type application/odata+json}},
>>> {"/data/Analytics.json" 0.600000 {type application/json}},
>>> {"/data/Analytics.jsonld" 0.500000 {type application/ld+json}},
>>> {"/data/Analytics.n3" 0.800000 {type text/n3}}, {"/data/Analytics.nt"
>>> 0.800000
>>> {type text/rdf+n3}}, {"/data/Analytics.ttl" 0.700000 {type text/turtle}},
>>> {"/data/Analytics.xml" 0.950000 {type application/rdf+xml}}
>>>
>> Given this Alternates-header: how can  a client figure out what those
>> representations look like (except for their media type)?
>>
>
> Your Web Browser (a client) understands text/html. A Browser and other
> HTTP clients apply the same content handling rules to other content types
> (e.g., those related to images, sound, and video etc..) .
>
>
>
>>  Link:
>>> <
>>> http://mementoarchive.lanl.gov/dbpedia/timegate/http://dbpedia.org/resour
>>> ce/Analytics>; rel="timegate"
>>> Location: http://dbpedia.org/data/Analytics.ttl
>>> Expires: Tue, 12 May 2015 16:01:06 GMT
>>> Cache-Control: max-age=604800
>>>
>> Best,
>>
>> Lars
>>
>
>
> --
> Regards,
>
> Kingsley Idehen
> Founder & CEO
> OpenLink Software
> Company Web: http://www.openlinksw.com
> Personal Weblog 1: http://kidehen.blogspot.com
> Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
> Twitter Profile: https://twitter.com/kidehen
> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen
> Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this
>
>
>


-- 
Paul Houle

*Applying Schemas for Natural Language Processing, Distributed Systems,
Classification and Text Mining and Data Lakes*

(607) 539 6254    paul.houle on Skype   ontology2@gmail.com
https://legalentityidentifier.info/lei/lookup
<http://legalentityidentifier.info/lei/lookup>