[foaf-dev] beyond foaf:mbox_sha1sum

Mischa Tuffield mischa.tuffield at garlik.com
Tue Dec 22 20:30:08 CET 2009


On 22 Dec 2009, at 18:31, Story Henry wrote:

> On 22 Dec 2009, at 13:36, Mischa Tuffield wrote:
>> Hi Henry, 
>> [snip]
>> On 22 Dec 2009, at 13:15, Story Henry wrote:
>>> [snip]
>>> When someone specifies a foaf:homepage, one can follow through the consequences of this: 
>>> - ask them if they want to make it their OpenID, and then give them instructions on how to do this, so that it can be verified. 
>> I make an attempt at this in the validator. If the FOAF document in question shares any IFPs with other foaf files in our in KB, we highlight them. You can see examples of this when you attempt to validate my foaf file : 
>> http://foaf.qdos.com/validator/uri?URI=http%3A%2F%2Fmmt.me.uk%2Ffoaf.rdf&VERBOSE=&PARSE=Parse+URI%3A+
> This took a bit of time to come up, but once it did I understood why...
> Very nice.
> I tried it of course on my own foaf file.
>   http://bblfish.net/people/henry/card

Ah yes, it does take a bit of time, the main culprit is the rdfs:seeAlso test, which goes out and checks whether the documents you rdfs:seeAlso are actually RDF documents.

Re: Your foaf file, if you would be so kind you could add an IFP to the bnode with foaf:name "Mischa...". You could add a _:bnodeX foaf:homepage <http://mmt.me.uk/> :) Thanks Henry :)

> Some things I noticed doing this:
> - the foaf spec has two properties for
>  http://xmlns.com/foaf/spec/#term_OnlineAccount
>    "May have properties: holds account holds account"
> - Test 3: Checking if PrimaryTopic is a foaf:Person:
>     Success. Your PrimaryTopic is a foaf:Person
>  should that be foaf:primaryTopic?
>  I suppose that is because the document is a PersonalProfileDocument. Normal foaf:Documents don't have that restriction.
>  If this is a foaf:PersonalProfileDocument, then I think it follows that the primary topic is a foaf:Person. So perhaps the thing to check would be if it follows from what is said in the doc, that it is not. (now that would be a bit of owl reasoning, but it should be ok, as you are limiting it to small documents)

The primaryTopic test is to check which subject URI or bnode for that matter is the "main" foaf:Person in the foaf file. This is due to the fact that I generally wouldn't trust statements made in a foaf:PersonProfileDocument that aren't directly associated with the primaryTopic. Basically, if there is a foaf:Document, with loads of foaf:Persons which one is the "main" person in this document.

> - Test 10: Checking if foaf:mbox_sha1sum have language tags
> ??? sha1_sums with language tags ???

Yeah crazy right? I found that this is way a rather common mistake, insofar as people adding languages tags to the root of their RDF/XML document, which in turn meant that the foaf:mbox_sha1sum's where rendered useless.

>> Sure, you can use our web-services, to help with this task: 
> Very nice services! http://foaf.qdos.com/

Thanks ...

>> [snip] and we do NO OWL stuff in the foaf.qdos world. All of the functionality has been implemented using a bunch of sparql queries. 
> Well one can do OWL using SPARQL CONSTRUCT queries of course :-) That is what I used the networked graph sail in Sesame for http://is.gd/5xySC in the Sommer Address Book.
> And you do use some of the inferencing I hope: such as if someone foaf:knows someone that both must be foaf:People....

Re: OWL, with my foaf.qdos.com hat on we do all of our "OWL based reasoning" at query time, we hope to have a larger KB full of foaf at some point in the future, and thought that we would end up generating far too many triples if we did created them at import time. Saying that was hoping to have a play with am implementation of OWL2 RL, if and when I get my hands on one, iirc Aberdeen have built one ...

And yes, a final big ++ to validators.


> What would be interesting also, in addition to your great work, would be to have semantic validators too that can work out clear oddities that follow from applying owl rules on a file.
> But thanks a lot, I have used it to fix a few problems in my foaf file. This is something to integrate in an Address Book too.
> Henry
> _______________________________________________
> foaf-dev mailing list
> foaf-dev at lists.foaf-project.org
> http://lists.foaf-project.org/mailman/listinfo/foaf-dev

Mischa Tuffield
Email: mischa.tuffield at garlik.com
Homepage - http://mmt.me.uk/
Garlik Limited, 2 Sheen Road, Richmond, TW9 1AE, UK
+44(0)20 8973 2465  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.foaf-project.org/pipermail/foaf-dev/attachments/20091222/d6f06e62/attachment.htm 

More information about the foaf-dev mailing list