[foaf-protocols] adding Ping to Simple WebID based Resource ACL

Kingsley Idehen kidehen at openlinksw.com
Sat May 22 17:44:52 CEST 2010

Nathan wrote:
> Story Henry wrote:
>> On 21 May 2010, at 19:57, Nathan wrote:
>>> Story Henry wrote:
>>>> On 21 May 2010, at 12:50, mike amundsen wrote:
>>>>>> You're now added to the ACL list.
>>>>> and it works like a champ!
>>>> We next need to add an automatic ping software to make this so that anyone can test it.
>>>> And perhaps even a form so we can remove ourselves from the group.
>>>> Idea: ping to add ourselves to the group. Use foaf+ssl to access a form to remove ourselves form the group. The people can see the difference between when they are in and when they are out.
>>> finally caught up & whilst I'd like a chance to implement some pinging - I don't think this is the right use case.
>> We need a simple use case to test ping, and see that it works. This is a very simple
>> one to get started.
>>> Why do we have to go to all this trouble, when in each of our foaf's we could simply add the triple
>>>  #me sioc:member_of <group> .
>>> and be done with it?
>> And how will Kingsley's server know you are now a member of the group? Unless you ping him, he'd have to crawl the whole internt - an impossible task - to find your file.
> it'd be in my foaf file, which he already needs to dereference and pull 
> in to auth my foaf+ssl, so he'd already have the info.. :)
> two sides to what I'm saying:
> If we're going to implement things I feel we should do the the 'Right 
> Way' and creating a server side API (however basic) for something that 
> doesn't actually need one, is, imho, not a good way to do the test 
> implementation of a new technology (the ping process & onto) - any 
> feedback which comes from it may not be expressly correct.
> On the flip side, obviously I'm keen to implement the ping process and 
> onto and get this model working, so all cool with doing it this way too, 
> just personally keeping it in mind that it's not the only option :)

Correct, not the only option. But it is the "embrace and extend" option 
re. Web 2.0 bridge building :-)
> Will give it a go over the weekend, quite sure it should be easy enough 
> to implement.
> aside: also worth considering that typical sparql update could be used 
> to add and remove the triples too, but that would mean somebody else 
> could remove your membership, so has draw backs.

If you give them "Delete" privs, then yes. With the relationship 
ontology you can be a little more granular about data access rules etc..

For now, we do the simple stuff, there is much more sophistication to 
come when we get to Data Access policies. In 1993 we were able to enable 
organizations like Xerox deliver policy based access to data over ODBC 
that respected corporate hierarchies etc. Could even stop a delete based 
on the ODBC compliant application or the IP address from which the app 
was being used + a raft of other factors. The rule book we built was 
based on an INI based representation of a EAV graph . We've implemented 
an RDF variant of our Rule book in the form of an OWL ontology a while 
back :-)

Once authentication is done, we'll have a lot of fun with Linked Data 
driven Data Access policies!!

> regardless, keen to see how this progresses - should be good!
> ps: no need to reply, just stating my thoughts on the matter
> Best,
> Nathan
> _______________________________________________
> foaf-protocols mailing list
> foaf-protocols at lists.foaf-project.org
> http://lists.foaf-project.org/mailman/listinfo/foaf-protocols



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

More information about the foaf-protocols mailing list