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

Re: design issue when dereferencing a foaf-profile with public key

From: Kingsley Idehen <kidehen@openlinksw.com>
Date: Thu, 29 Sep 2011 18:59:35 -0400
Message-ID: <4E84F857.7090801@openlinksw.com>
To: public-xg-webid@w3.org
On 9/29/11 4:41 PM, Jürgen Jakobitsch wrote:
> hi again,
>
> that's exactly what i was talking about, there's no path directed from the foaf:Person
> into the direction of the RSAPublicKey, only one that is directed from RSAPublicKey
> to the foaf:Person : that's why i was asking if a predicate like hasRSAPublicKey would help,
> please also note that i sayd below that "there's no statement about the foaf:Person indicating where to find the RSAPublicKey"
>
> i removed my public key from http://www.turnguard.com/turnguard and put it elsewhere. if i gave you the url (uri) of my
> public key now, it wouldn't test what i'm talking about, because when logging in using a certificate, the only uri i have
> is the one from the foaf-profile.
>
> if you find my public key, congrats and i owe you a drink.
>
> nevertheless i would have to consider this a trick on application level. the webID spec can hardly say that
> implementors should have a lod cache at hand..
>
> please also note that i cannot log in now @ fcns.eu for example, which is no surprise.

When you say: "it's valid Linked Data" one has to assume there is a 
connecting relation that's follow-your-nose friendly, courtesy of 
indirection prowess delivered by de-referencable URI based object 
(resource) identifiers.

You can craft a 3-tuple (triple) that includes URIs in a numbers of 
ways. Thus, I am assuming that you are factoring in:

1. RDF triples
2. <link/> based triples courtesy of Web Linking patterns,

with regards to the fundamental 3-tuple based relation:

<http://kingsley.idehen.net/dataspace/person/kidehen#cert1> 
<http://www.w3.org/ns/auth/cert#identity> 
<http://kingsley.idehen.net/dataspace/person/kidehen#this> .

Kingsley

>
> enjoy
> http://www.turnguard.com/turnguard
>
> ----- Original Message -----
> From: "Kingsley Idehen"<kidehen@openlinksw.com>
> To: public-xg-webid@w3.org
> Sent: Thursday, September 29, 2011 9:46:54 PM
> Subject: Re: design issue when dereferencing a foaf-profile with public key
>
> On 9/29/11 3:04 PM, Jürgen Jakobitsch wrote:
>> hi kingsley,
>>
>> thanks for your reply,
>> may i ask, how you would find the publickey of the following fictional foaf-profile :
>>
>> <foaf:Person rdf:about="http://www.someuri.org/card#me">
>> ...
>> </foaf:Person>
>>
>> <rsa:RSAPublicKey rdf:about="http://www.public-keys.net/2342">
>>     <cert:identity rdf:resource="http://www.someuri.org/card#me"/>
>> </rsa:RSAPublicKey>
>>
>> when these resources are dereferenceable under their respective uri
>> and my certificate states that the subject's UID is http://www.someuri.org/card#me
>> and there's no statement about the foaf:Person indicating where to find
>> the RSAPublicKey.
>>
>> i cannot quite believe you crawl the whole lod cloud to find a statement
>> ?x cert:identity http://www.someuri.org/card#me
>>
>> maybe i just have a knot somewhere in my brain...
>>
>> any pointer very welcome
>> wkr http://www.turnguard.com/turnguard
> Why don't you post the URLs of the two resources. Then lets see if I can
> use SPARQL to get you a match.
>
> If there is a path connecting the two resources hosting the relations,
> the Virtuoso's SPARQL can find it.
>
> Also note, we have a living agent on the Web called URIBurner (curated
> by its users since 2007), we host a massive LOD Cloud Cache (29 Billion+
> Triples), these things are all hooked together is Webby ways.
>
> So you can send me the URLs, or hopefully I've added some clarity to the
> nascent puzzle :-)
>
> Kingsley
>>
>> ----- Original Message -----
>> From: "Kingsley Idehen"<kidehen@openlinksw.com>
>> To: public-xg-webid@w3.org
>> Sent: Thursday, September 29, 2011 8:34:36 PM
>> Subject: Re: design issue when dereferencing a foaf-profile with public key
>>
>> On 9/29/11 2:06 PM, Jürgen Jakobitsch wrote:
>>> hi all,
>>>
>>> i have a question concerning linked data principles and a dereferencing a foaf-profile
>>> with a public key.
>>>
>>> currently it is apparently necessary that two (=2) resources are dereferenceable under one (=1) (the foaf:Persons's) uri.
>>> that is because there's no predicate linking from a foaf:Person to a RSAPublicKey. I could have a RSAPublicKey
>>> available at some-uri stating that it's cert#identity is some remote resource, which would be totally legal
>>> linked data, but when using the foaf uri from a certificate there's no chance i find the coresponding RSAPublicKey.
>> You assume that there are no WebID authentication/verification protocol
>> implementations that also include follow-your-nose crawling :-)
>>
>> FYI -- that's integral to Virtuoso's WebID implementation. It even
>> includes reasoning and transitive closures at LOD scales.
>>> 1. has this issue already been discussed?
>> Yes, there was a thread between Peter Williams and I about this. He
>> raised this matter way back, so to speak.
>>
>>> 2. is it not considered an issue but simply the way it is?
>>> 3. would a predicate like "hasPublicKey" improve things?
>>>
>>> any comment or opinion really appreciated
>>> wkr http://www.turnguard.com/turnguard
>>>
>>>
>


-- 

Regards,

Kingsley Idehen	
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen








Received on Thursday, 29 September 2011 22:59:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 29 September 2011 22:59:59 GMT