W3C home > Mailing lists > Public > public-xg-webid@w3.org > November 2011

Re: long term webid of peter, contrary to rumour

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Mon, 14 Nov 2011 07:11:36 -0500
Message-ID: <4EC10578.3060009@openlinksw.com>
To: public-xg-webid@w3.org
On 11/13/11 5:17 PM, Henry Story wrote:
>
> On 13 Nov 2011, at 22:48, Kingsley Idehen wrote:
>
>> On 11/13/11 3:53 PM, Henry Story wrote:
>>>
>>> On 13 Nov 2011, at 21:31, Kingsley Idehen wrote:
>>>
>>>> On 11/13/11 4:48 AM, Henry Story wrote:
>>>>> On 13 Nov 2011, at 01:52, Peter Williams wrote:
>>>>>
>>>>>> atyorkporc.wordpress.com <http://yorkporc.wordpress.com/>ive 
>>>>>> hosted on the blog's front page the site's contact page 
>>>>>> (fromwordpress.com <http://wordpress.com/>). It has in HTML the 
>>>>>> kind of information normally shown in a foaf card. it has my long 
>>>>>> term webid, hosted on an opera unite endpoint. Its not a foaf 
>>>>>> card like others and neither is the endpoint (being only 
>>>>>> available when I am online).
>>>>>
>>>>> That's ok. As the spec points out ( 
>>>>> http://www.w3.org/2005/Incubator/webid/spec/ ) all that is 
>>>>> required for WebID is the publication of the public key at that 
>>>>> endpoint with some RDF markup. (btw, we should perhaps add a link 
>>>>> to the W3C how to on publishing multiple formats in a content 
>>>>> negated format)
>>>>
>>>> Is it about publication of a public key with RDF markup? Is that 
>>>> the narrative? If it is, then be up front about it as I am tired of 
>>>> cycling this RDF wagon re. the problem it introduces, unnecessarily.
>>>>
>>>> If this is an RDF only solution, say so, and stick to it. Then live 
>>>> with the consequences.
>>>
>>> Is there an issue you have with the spec? If so please tell us.
>>
>> I have an issue with narratives the end up with RDF as being 
>> inextricable re. WebID and its verification protocol. That's what I 
>> have an issue with. If the spec toes that line, then I have a problem 
>> with the spec. If the spec is RDF specific then qualify the whole 
>> thing as RDF based WebID, nice and simple.
>
> We have RDF/XML, Turtle, RDFa markup in html. Where is RDF/XML 
> inextricably linked?

I didn't mention RDF/XML, I referred to RDF. As you can see, the letters 
R-D-F are prone to confusion, all the time. And when it isn't confusion 
you end up with gut-reaction style political reactions. This is the sad 
nature of RDF, unfortunately.


> We speak about the model, and we show the serialisations that are 
> widely accepted.

You think you are speaking about a model, but you aren't. You are 
speaking about RDF syntaxes. We have a directed graph model that holds 
claims that can be matched to claims in an x.509 based security token 
via the semantics of its relations. We can talk about WebID without 
implementation details like the syntaxes for directed graph representation.

RDF is but one set of syntaxes for representing directed graphs. It just 
isn't much more than that. RDFs and OWL semantics aren't inextricably 
bound to RDF syntax either.

>
> There has to be a way of telling in follow your nose like manner how 
> to get the graph, which does not rely on things like: if the service 
> is called Facebook, then do this, but if it is Twitter then do that, 
> and if is some other site then do that.

In our solution we don't do anything like that. We just use the 
Fingerprint (hash of claims) as an option for determining the mirror of 
the claims held in the x.509 based security token . The only issue with 
the current incarnation of our solution is the fact that its based on us 
writing drivers for the Facebook, Twitter, LinkedIn, and AtomPub 
protocols. The next incarnation will simply build on that and produce 
ProxyURIs, once in place others relying agents won't have to implement 
their own drivers since follow-your-nose will just work since the URIs 
will resolve to a navigable directed graph serialized in a variety for 
graph representation formats (RDF and non RDF family).

>
> Or how do you think we should currently work with Peter William's 
> profile?

Peter is seeking a way to place the mirror of his x.509 claims into a 
WordPress post. The problem he has is that like most Web 2.0 solutions, 
Wordpress is very restrictive, so he can't put (X)HTML+RDFa in there.

> Should we perhaps add something to the spec that says if the URL is
>
> $ curl -i http://home.homepw2.operaunite.com/webserver/content/
> HTTP/1.1 503 Service Unavailable
> Content-type: text/html
> Connection: close
> Server: UniteProxy/0.2.5
>
> <html><frameset cols="100%"><frame 
> src="http://unite.opera.com/general/noservice/homepw2/home/" 
> /></frameset></html>
>
> then we should go to http://yorkporc.wordpress.com/ and read the 
> public key there by searching for the "RSA Public Key" string
> and then finding the key by guessing that that's probably a modulus 
> because it looks like one?

No.

>
> And even if we were to find the public key there, we would find that 
> the webid does not point to the right place but to a different 
> document that is unavailable. But perhaps that's acceptable because 
> the spec should say that if its Peter William's site we should have an 
> exception.
>
> Clearly you are not going to defend such a position.

Of course not. But I am interested in pragmatic ways of publishing WebID 
related claims to Blog posts. I know this is critical to mass adoption 
of WebID.

> But currently I don't see how Peter Williams can claim that he even 
> has a WebID, not in any meaningful way related to this group's work.

Peter is seeking a pragmatic solution devoid of religion and politics.

Kingsley
>
>>
>>>
>>> Most implementations I know of now are working with RDF 
>>> serialisations, so those are the ones we should be sticking by 
>>> initially, as we did from the very start.
>>
>> Again, what on earth does that mean? That there will be a narrative 
>> utterly laced with that bias? Again, there's nothing wrong with 
>> saying: this is RDF based WebID etc.. That's better that pretending 
>> it isn't be it via spec or narrative.
>
> So what do you want the spec to say?
>
>>
>>> Those serialisations are well documented and clearly specified.
>> So?
>>
>> Kingsley
>>>
>>>>
>>>>>
>>>>> I see that your WebID Profile Page - as it is called in the spec 
>>>>> section 2.3 - is in html. So I guess it's meant to be parsed as RDFa.
>>>>> The RDFa http://www.w3.org/2007/08/pyRdfa/ validation service 
>>>>> seems to only return a few URLs for your page.
>>>>>
>>>>> I don't see that you have specified any of the cert or rsa 
>>>>> namespaces so if you want to turn that into a WebID you do need to 
>>>>> follow the explanation here
>>>>>
>>>>> http://www.w3.org/2005/Incubator/webid/spec/#rdfa-html-notation
>>>>>
>>>>> If you find problems or unclarities in any aspect of the spec, 
>>>>> please explain which part of the text is unclear, and what wording 
>>>>> you suggest would help improve it.
>>>>
>>>> You are talking syntax again. Can WebID not be discussed 
>>>> conceptually without syntax specificity? Is this impossible?
>>>
>>> I am talking spec.
>>>
>>> In the future when semantics is more clearly and widely understood 
>>> then one will no longer need to mentions syntax. But at present that 
>>> is not the case. The document is an evolving one.
>>>
>>> Henry
>>>
>>>>
>>>> Kingsley
>>>>>
>>>>> Henry
>>>>>
>>>>>>
>>>>>> I happen to enforce more privacy than perhaps do most consumers 
>>>>>> (being a security type engineer who is experimenting with 
>>>>>> semantic web ideas, as they evolve). I know some folks want foaf 
>>>>>> cards as public data, cacheable by search engines and others 
>>>>>> maintaining huge triple stores. I dont. I want to assert my 
>>>>>> privacy expectations (becuase in the US, one has no rights until 
>>>>>> they are asserted - this being the way that social laws on 
>>>>>> privacy happen to be structured.) And I do this in ways that may 
>>>>>> not meet the idealized, academically-normalized semantic web 
>>>>>> concept, which assumes you have public documents that anyone can 
>>>>>> browse, cache, cite from, etc.
>>>>>>
>>>>>> Now
>>>>>>
>>>>>> does anyone know HOW to make a wordpress contact page embed either
>>>>>>
>>>>>> (1) an XML stream, coding up the xml-serializion of a foaf card 
>>>>>> (e.g. the output of foaf.me), or
>>>>>> (2) the RDFa of a foaf card.
>>>>>>
>>>>>> I cannot make either work.
>>>>>>
>>>>>> I seem to remember that one was supposed to be able to add  "XML 
>>>>>> data island" as a subelement in HTML (since XHTML is just an XML 
>>>>>> vocab). One is supposed to be able to embed RDFa tags in HTML 
>>>>>> elements too. Unfortunately, the wordpress site strips out what I 
>>>>>> add (refusing to host them). Ill guess that these are default 
>>>>>> safety options, that can be removed by those who know how.
>>>>>>
>>>>>> My wordpress blogsite is hosted by wordpress cloud service, not 
>>>>>> by me on a server (im not sufficient competent to run a 
>>>>>> production server). So, there may be less ability to change the 
>>>>>> configuration to offer RDFa and XML embedding, than a site one 
>>>>>> hosts on a private server.
>>>>>
>>>>> Social Web Architect
>>>>> http://bblfish.net/
>>>>>
>>>>
>>>>
>>>> -- 
>>>>
>>>> Regards,
>>>>
>>>> Kingsley Idehen	
>>>> President&  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
>>>>
>>>>
>>>>
>>>>
>>>
>>> Social Web Architect
>>> http://bblfish.net/
>>>
>>
>>
>> -- 
>>
>> Regards,
>>
>> Kingsley Idehen	
>> President&  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
>>
>>
>>
>>
>
> Social Web Architect
> http://bblfish.net/
>


-- 

Regards,

Kingsley Idehen	
President&  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








Received on Monday, 14 November 2011 12:12:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 14 November 2011 12:12:15 GMT