Re: Linked Data and Semantic Web CoolURIs, 303 redirects and Page Rank.

On 7/23/14 5:15 PM, Michael Smethurst wrote:
>
> On 23/07/2014 21:49, "Kingsley Idehen" <kidehen@openlinksw.com> wrote:
>
>> On 7/23/14 3:40 PM, Michael Smethurst wrote:
>>> Hi Kingsley
>>>
>>> Very definitely starting to feel like deja vu...
>>>
>>> On 23/07/2014 20:18, "Kingsley Idehen"<kidehen@openlinksw.com>  wrote:
>>>
>>>>> On 7/23/14 2:05 PM, Michael Smethurst wrote:
>>>>>>> For internal usage it's all probably fine. But I still think it's a
>>>>>>> pattern that shouldn't be generally encouraged.
>>>>> Its a "horses for courses" matter:-)
>>>>>
>>>>> If you choose to use hashless HTTP URIs in regards to entity
>>>> denotation,
>>>>> you have to make the extra investment required (via 303 heuristics)
>>>> for
>>>>> entity disambiguation [1].
>>> My only point is: if you don't conflate "I can't send that" (303) with
>>> "what flavour would you like" (conneg) you don't have to invest in more
>>> servers
>>>
>>>>> Note, there are changes to HTTP that also reduce some of the confusion
>>>>> in this realm. For instance the use "Content-Location:" response
>>>> headers
>>>>> to aid disambiguation [2].
>>> We do use content location for the (information) resource /
>>> representation
>>> split but that's REST not 303 semantics
>>>
>>> michael
>> There is only one kind of relation semantics in play here, and its the
>> semantics of denotation and connotation [1][2].
> Tho derrida didn't have to pay for servers :-/
>
>> HTTP URIs denote things.
> Which can't be served (303)

No, HTTP URIs simply denote things (entities). It has nothing to do with 
being served etc..

>
>> HTTP URLs denote documents comprised of connotation bearing content.
> Which can be served in assorted representations (conneg (+ content
> location))

No, HTTP URLs are a kind of HTTP URI that denote Web Documents. Put 
differently, HTTP URLs are for all intents an purposes a colloquialism 
for HTTP URIs that focuses on Web Documents, a particular entity type 
i.e., entities that are instances of the Classes denoted by the URIs: 
<http://xmlns.com/foaf/0.1/Document>, 
<http://purl.org/ontology/bibo/Document>, 
<http://purl.org/dc/terms/BibliographicResource> etc..

The very same analogy applies to WebIDs which are HTTP URIs that denote 
Agents i.e., entities that are instances of the Class denoted by the 
URI: <http://xmlns.com/foaf/0.1/Agent> .
>
> Think the last time we had this conversation we broke the twitter scroll
> bar and agreed to disagree. Or at worst misunderstand :-)

Long discussions aren't necessarily bad, they can also unravel insights 
that are sometimes overlooked :-)

BTW -- one can also deconstruct this issue a different way, starting 
with HTTP URI/URLs that denote Documents. It goes something like this:

1. You have a RDF document (comprised of RDF/XML content) denoted by the 
HTTP URI/URL <http://www.bbc.co.uk/programmes/b006mw1h.rdf>
2. The document above describes an entity denoted by the HTTP URI 
<http://www.bbc.co.uk/programmes/b006mw1h#programme> .

We arrive at the same place (as illustrated by the Vapour links I shared).

My only issue with the BBC programmes URIs right now is that 
<http://www.bbc.co.uk/programmes/b006mw1h> doesn't make its association 
with <http://www.bbc.co.uk/programmes/b006mw1h#programme> discoverable 
to RDF user agents. That's where Microdata, RDFa, <link/>, "Link:" come 
into play i.e., they provide vehicles for exposing the missing relation 
(association, connection, relationship property/predicate etc..).

Also note:

curl -IH "Accept: application/rdf+xml" 
http://www.bbc.co.uk/programmes/b006mw1h
HTTP/1.1 200 OK
Server: Apache
Content-Type: text/html; charset=utf-8

curl -IH "Accept: application/rdf+xml" 
http://www.bbc.co.uk/programmes/b006mw1h.rdf
HTTP/1.1 200 OK
Server: Apache
Content-Type: application/rdf+xml

Which reinforces my point re. missing relation to aid RDF user agents. 
Simply tacking on ".rdf" to the end of URLs is way too brittle, when a 
relation (describes, describedby etc..)  would do much better via RDFa, 
Microdata, <link/>, "Link:" etc..




Kingsley
>
> michael
>> In regards, to the current BBC programmes URIs, if you incorporate RDFa,
>> <link/>, or "Link:" based relations, disambiguation without 303's or
>> content negotiation is possible. RDF user agents (for example) will be
>> able to make sense of the relations that that collective describe
>> documents about programmes and actual programmes.
>>
>> Links:
>>
>> [1] http://bit.ly/what-does-this-bbc-programmes-uri-denote -- Vapour
>> using RDF semantics discern what
>> <http://www.bbc.co.uk/programmes/b006mw1h> denotes and connotes
>>
>> [2] http://bit.ly/what-does-this-bbc-programmes-doc-url-denote -- ditto
>> but targeting <http://www.bbc.co.uk/programmes/b006mw1h.rdf> .
>>
>> -- 
>> 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
>>
>>
>
>


-- 
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