Re: Important Question re. WebID Verifiers & Linked Data

On 12/21/11 11:27 AM, Henry Story wrote:
> On 21 Dec 2011, at 14:58, Kingsley Idehen wrote:
>
>> Henry,
>>
>> Is a WebID verifier supposed to be a Linked Data consumer ?
>> Is a WebID verifier supposed to be an RDF consumer?
> In the spec it clearly is an RDF consumer, yes.

So WebID is at best RDF based. But not Linked Data based. Maybe, RDF 
based Linked Data driven sorta. Then you hit the other issue re. AWWW 
since you are also tossing aside the very essence of URIs. Again, a user 
agent doesn't have a say in the matter re. what resources a URI resolves 
too. A smart user agent can negotiate resource representation, but 
that's as far as it goes.



>   All our implementations have followed
> this since the beginning.

We haven't and never will!

>   I am not sure if the Verifier itself needs to be a Linked
> data engine, but it is best for authorisation that it is. That's where things get
> interesting.

You betcha!
>
>> Please understand that RDF != Linked Data. It's just one of the options for creating and publishing Linked Data.
> I think it would be very nice to have a formal spec on what Linked Data is. We do have a few for
> RDF.

Linked Data is very very simple. It is about URIs that resolve to 
resources that bear the representation of the URI's referent. Simple and 
unambiguous. The description takes the form of an eav/spo graph 
pictorial. Actual representation of the resource bearing the eav/spo 
pictorial is negotiable.

This is what I've been trying to communicate to you for ages re. these 
distracting arguments about RDFa vs Microdata vs RDF/XML vs any other 
eav/spo based resources that serves as an entity/object/thing descriptor.

None of this stuff is new, its how stuff has worked for eons. The clever 
innovation by TimBL is the invention of URIs and the application of URI 
abstraction to data access and data representation in a manner that 
transcends platforms while also leveraging network scale e.g., InterWeb. 
Then on top of that he made HTTP which separates data access protocols 
from data representation. Basically, he has used Hyperlinks to deliver 
the deep power power of pointers in a very elegant manner.

Big problem is that the likes of Java and other staples from the OO 
world discouraged people from learning about pointers. Thus, we have a 
generation of developers that don't understand name / address 
disambiguation, indirection, address-of etc..

>
>> In addition:
>>
>> Is a WebID verifier supposed to be an HTTP client?
> If by supposed to be, then yes. I think all of those we know of are. That why it is a Web Identity, not an SMTP identity
> for example.

Well HTTP != Web. It's a low cost protocol delivering the virtues 
outlined above. WebID is supposed to be about URIs, the "Web" part has 
nothing to do with HTTP bar low cost benefits it delivers.
>
>> Is a WebID verifier supposed to be able to leverage HTTP content negotiation?
> We are just discussing that. I am in favour, but the point is to look at the security issues if any.

Don't use "security" in this manner. I find it unbecoming of you, 
seriously now. You wouldn't tolerate that kind of FUD from anyone else.

A URI isn't implicitly secure or insecure. It's a URI.

>
>> Right now, WebID verifiers are very inconsistent re. the above, thus I encourage we bring clarity to this very important manner.
> Which ones are inconsistent?

Yours and any others that are break AWWW. Put differently, any verifier 
that discriminates against hash or slash URIs is broken. Again, this 
matter is none negotiable. You either honor AWWW or you don't.

>
>> Remember, if we are going to tout WebID as leveraging Web Architecture, we have to actually comply with said architecture.
> I don't see that we are not. What in the spec is not compliant with web architecture?

Please digest my comments. I've been crystal clear.

You are currently breaking AWWW by overreaching.

For now you really have RDF based WebID for HTTP. Its based on one level 
of indirection, and it doesn't really honor Name / Address 
disambiguation of Linked Data. It also doesn't honor URI abstraction.
>
> Henry
>


-- 

Regards,

Kingsley Idehen	
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen