[OpenID] cryptographics web of trust

[Story Henry]

On 7 Sep 2007, at 20:48, Peter Williams wrote:
>
>> -----Original Message-----
>> From: Story Henry [mailto:henry.story at bblfish.net]
>> Sent: Wednesday, August 29, 2007 4:18 AM
>> To: Peter Williams
>> Cc: OpenID General
>> Subject: Re: [OpenID] cryptographics web of trust
>>
>> On 28 Aug 2007, at 22:22, Peter Williams wrote:
>>
>>> Don't laugh.  I'm naturally slow. I half got my blog site to work,
>>> and got a first RDF FOAF file. I may even have a persistent URI.
>>>
>>> http://yorkporc.spaces.live.com/blog/cns!5061D4609325B60!148.entry
>>>
>>
>> Nice! There's nothing like doing to understand!
>
> The blog provider I'm using (MSN Spaces) is very controlling. I cannot
> get it to allow link tags of type meta.
>
> I know you have a nice scheme where the HTTP can act as a layer 6
> presentation negotiation protocol - letting the peers negotiate the
> serialization format of the RDF model, while on the wire. But, this
> would require the consumer to run and operate a webserver, rather than
> merely edit some blog site and place a link tag.

yes. this is using simple HTTP Content negotiation
http://www.w3.org/Protocols/HTTP/1.1/rfc2616bis/draft-lafon- 
rfc2616bis-03.html#content.negotiation


But of course people are most likely to ask for the html  
representation first so having a link
from the html representation to the rdf representation is a good idea.

> It is accepted TODAY that one denote FOAF metadata associated with an
> HTML file using these conventions and file extensions? (from a 5 years
> old book?)
>
>
> <link rel="meta" type="application/rdf+xml" href="my-foaf- 
> file.xrdf" />

Due to limitations of the link relation (it does not allow one to  
specify as in atom the type of the link [1])
You need to add the title="FOAF" in html as a hint that one is  
linking to a foaf file.
This is shown here: http://www.w3.org/QA/Tips/use-links

This will be solved in different ways by xhmtl 5, and by RDFa [2],  
but for the moment this is an acceptable solution.

> (Obviously, the HTML reader has to be metadata aware to leverage those
> signals, and know that the stream resulting from following the link
> should have: application/rdf+xml serialization.

Yes, not unlike the openid link. Just as with with an openid client,  
the consumer need
not be fully html aware...

>
> Finally, be proud.
>
> I finally got my federation server (which is JDBC and JDBC/ODBCbridge
> capable) to pull attributes from a FOAF file, when doing WebSSO.
> http://www.intellidimension.com/ was the mature application server,  
> for
> the semweb.

Intellidimension has very good reviews. Sadly it only works on Windows.

> Now I need some code that can answer: if all humans are linked by 6
> know-relations, and I have a model of 1,000,000 foaf files, how do I
> compute FOAFishly (rather than Euler, for now) the actually relations
> that connect X to Y. I assume this has been done, lots of times, by
> spidering type solutions.

I think there are going to be many ways of calculating this, some  
more efficient than others. Of course with millions of people this  
efficiency is going to be a big issue.

But dealing with millions of foaf files would be overkill. Though I  
can imagine larger aggregators building up large databases of foaf  
related information, I don't think this is the only useful way of  
proceeding. And I also don't believe it is the most reliable.  It  
will also not be complete because people will be limiting who can see  
their complete foaf profile for security reasons. [3]

Foaf can also be very useful if you limit yourself to your friends  
and their friends, as LinkedIn does.

Henry

[1] http://www.atompub.org/rfc4287.html#rel_attribute
[2] http://www.w3.org/2006/07/SWD/RDFa/primer/
[3] http://blogs.sun.com/bblfish/entry/my_bloomin_friends
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2429 bytes
Desc: not available
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20070909/24271ad2/attachment-0002.bin>