On 18 July 2010 18:18, Stéphane Corlosquet <[email protected]> wrote:

> Just to clarify, you need to make sure the resources you're linking with
> owl:sameAs are of type foaf:Agent (or foaf:Person). You shouldn't directly
> link foaf:OnlineAccount since these have clearly different properties, e.g.
> account creation date, user name, content created with this account, the
> site it is associated with. This applies to the owl:sameAs property which is
> well deployed and understood by semantic web tools. In the end, to put this
> in more human friendly form: it's like saying that 2 accounts are owned by
> the same person, as opposed to saying that the 2 accounts are the same.


Very good point.

I was more illustrating the philosophy of linking two FOAFs together using a
hyperlink, rather than talking about type checking and binding to
foaf:Person ( I think that part's relatively easy to get right at
implementation time, once you've grokked the concept of linking )


>
> Of course, end users won't have to worry about this subtlety and it's easy
> to built this in the system. You can easily do this by using hash URIs for
> your foaf:Agent aka WebID, that's what most people and systems do, timbl
> uses card#i, Drupal uses user/23#me, etc.
>

Absolutely, this a a great pattern.  One step a time tho ...

I'd actually love to see drupal 7 and gnu social link up in this way too ...
that's something quite exciting that you can do in one line of code.  Now
that's what I call interoperability!


>
> Steph.
>
> On Sat, Jul 17, 2010 at 6:57 PM, Melvin Carvalho <[email protected]
> > wrote:
>
>> The topic came up from mattl on IRC of linking two accounts together, say,
>> for example on status.net and libre.fm.
>>
>> I discussed the FOAF perspective with timbl and it seems there is a
>> relatively straightforward solution to this:
>>
>> 1a. In libre.fm foaf is at ( say libre.fm/mattl )
>> 1b. In status.net foaf is at ( say status.net/mattl )
>>
>> 2. You just need a add a sameAs link between the two in your FOAF.
>> *
>> <http://libre.fm/mattl> sameAs <http://status.net/mattl>
>> *
>> And you're done.
>>
>>
>> Web 3.0 enabled search engines, visualization tools (e.g. tabulator), and
>> inferencers, should be able to understand this and make semantic inferences
>> appropriately.  It also makes federated single sign more straightforward.
>>
>> "a hub sort of hub approach" it was described by timbl ... hope that
>> helps!
>>
>>
>

Reply via email to