[foaf-dev] What is the best way to edit a FOAF profile?

Stefano Bortoli bortoli at disi.unitn.it
Thu May 5 16:53:05 CEST 2011


Dan Brickley wrote:
> On 5 May 2011 15:02, Leigh Dodds <leigh.dodds at talis.com> wrote:
>   
>> Hi,
>>
>> On 5 May 2011 11:12, Stefano Bortoli <bortoli at disi.unitn.it> wrote:
>>     
>>> Have you checked out the foaf-o-matic?  http://www.foaf-o-matic.org/
>>>       
>
> Yep, foaf-o-matic / foaf-a-matic, very subtle spelling difference...
>
>   
Foaf-o-matic in its first version was my master thesis project, and I 
tried to show it to Dan when we met at SWAP 2008 in Rome but we had 
connection problems :-(
The current version was developed in the context of the Okkam project 
and allows the integration of URIs issued by the Entity Name System we 
developed. It is not mandatory thou. ;-) Btw, foaf-a-matic allows also 
to import directly single or lists of friends from others FOAF profiles, 
and further allow to import directly vCard, for personal and friends 
information.

The name foaf-O-matic was actually inspired by the work of Leigh. We 
wanted just to remark the relation with Okkam with the big O. :-)
>> I wasn't aware of that. Looks really neat!
>>
>> A reference to the (hopelessly outdated, I know) original wouldn't
>> have gone amiss though :(
>>     
>
> How hopeless is hopeless? I remember once I tried customising a
> version for foaf site but never got some part working (xslt?).
>   
I guess foaf-a-matic is based on a Java library and Jena for the 
processing of the RDF. Also foaf-O-matic is based on Jena.
> Would you consider putting it in Github where folk could work on
> updating it? The naming vocab has improved a little since
> (PortableContacts allignment etc).
>
> The big problem with all these tools was always: what to do after the
> file has been generated. Asking end users for FTP passwords is not a
> good plan, ... (a) they don't usually have them (b) they don't usually
> know about them anyway (c) it's not very good security practice to
> encourage users to give such info away... So we always had this
> awkward limit with self-hosted FOAF, ... that we could generate a file
> for people but getting it uploaded, ... and linked, was tough.
>   
You are right. However, people that can actually publish their profile 
on a web server should not fear to use foaf-O-matic. It is an academic 
project and I can ensure you we are not interested in storing passwords 
anywhere. :-) if I wasn't ashamed of its quality, I would even point you 
the code to check. ;-)
> In recent years, rise of OAuth gives some hope. And WebID is somehow
> related. But neither give us the protocol for uploading files. That's
> more AtomPub's job, and AtomPub doesn't seem yet to be wildly
> successful, at least it isn't well deployed with OAuth. If it was, ...
> we could upload FOAF files either as attachments to 'blog posts', or
> in RDFa form, directly as blog posts. The issue then becomes how does
> a dated blog post get hyperlinked from someone's homepage. But that's
> a much smaller problem at least, ... and there are other discovery
> mechanisms possible (eg. hyperlinks from the post to directories that
> check http referers...).
>
>   
Why would one need yet another protocol to upload a foaf profile? to 
perform atomic updates of informations? :-\ Interesting, but the atomic 
post would have to be 'RDF enabled', as the update could possibly be 
referring to the primary topic, or to one of the known friends. right? 
Collection seems to be quite limiting in that sense.

However, updating friends information within the RDF graph of a FOAF 
profile requires friends to be explicitly identified and reachable 'from 
outside' the graph. The blank node approach relying on the value of 
sha1_mbox does not seem practical.

> The self-published FOAF model was really a bootstrapping idea that got
> the idea out there, rather than something that worked very well. 
>
>   
yes, it was an exciting experiment. Peter Mika even won a Semantic Web 
Challenge integrating those profiles. :-)

saluti,
Stefano

-- 
Stefano Bortoli, University of Trento
Department of Engineering and Information Science (DISI)
WWW: http://disi.unitn.it/~bortoli
Skype: stefano.bortoli
OKKAM id: http://www.okkam.org/ens/id8af7c50f-f072-4384-905b-03875c341863



More information about the foaf-dev mailing list