Re: Another attempt at 'cascading DRs'

Hi, Kai.

> [snip]
> 
>> In theory, we can (operationally) enforce a mechanism able to 
>> determine whether a DR scope S1 is more *specific* than 
>> another DR scope S2 wrt a given resource. For this purpose, 
>> we have just to check whether the set of URIs denoted by the 
>> regex corresponding S1 includes the set of URIs denoted by 
>> the regex corresponding to S2. Which is exactly what you said, Kai.
> 
> Correct.  I suggest to use this an indicator or importance (which what
> cardinality means in this context, right?).

Hmm... I'm not sure we are using that term with the same meaning. In 
this context, a cardinality restriction denotes the minimum/maximum 
number of occurrences of a given property in describing a given entity, 
or a given set of entities.

So, if we wish to state that a (set of) resource(s) cannot be, at the 
same time, safe and unsafe for children, we must state that property 
childSafe can occur at most 1 time, i.e., such property should have 
mincardinality=0 and maxcardinality=1.

> [snip] 
> 
>> Note that this solution requires to check the definition of 
>> property childSafe in the corresponding RDF schema. 
>> Operatively this means:
>>
>> 1. Retrieve the RDF/OWL schema defining childSafe 2. If 
>> there's no restriction on the cardinality of childSafe, both 
>> descriptors apply; otherwise, 3. If the scope definition of 
>> one of the DRs is more specific than the other wrt the 
>> resource's URI, the former overrides the latter; otherwise, 
>> 4. If the two scope definitions are equivalent/incomparable... what? 
>> (ignore both descriptors?)
> 
> I think you have described it quite well.
> Using the specificity of the URI we can determine which descriptor
> applies more.
> 
> You bring up an interesting point with 4.
> I would suggest if the scope is equal, it is recognized that both
> apply...somewhere within the scope.
> That is not a bad thing.  All it says that we are lacking knowledge, at
> that instance of precicely /*where*/ the differences occur.
> This may lead to resource retrieval or not, depending on the context in
> which the user is at that moment.
> 
> If, for example, the descriptors that there is childsafe content and
> childunsafe content, a mother may have set her FOSI client to block the
> entire scope.
> Alternatively, if the descriptors say that some of the T-shirts for sale
> are blue and some a red, the potential shopper will still retrieve
> resources within that scope.

But, in our example, this wouldn't match with the cardinality 
restriction concerning childSafe, since it would imply two occurrences 
of such property. I can see here the following possible options:
1. ignore the descriptors concerning property childSafe
2. decide the prevailing one based on the associated value

In case of option 2, who can decide whether "true" prevails over "false" 
or vice versa? The individual/organisation which defined property 
childSafe? The individual/organisation author of the DRs? The end user? 
Probably the last one, following your considerations. But how can we 
enforce this?

Also, it must be taken into account who are the authors of such 
contradictory statements. Suppose the following DRs:
DR1: example.org/foo -> childSafe=true
DR2: example.org/foo -> childSafe=false

Now, if DR1 and DR2 have different authors, whether the former prevails 
over the latter may depend on how much I trust their authors (as Phil 
pointed out in one of his previous emails), or also based only on the 
value of childSafe (as in your example above). By contrast, if DR1 and 
DR2 have the same author, what does this mean? Different authors may 
have different opinions, but the same author cannot say that the same 
set of resources are both safe and unsafe for children. Such 
contradiction might imply that the author made a mistake. So, in this 
case, it may make sense to ignore such descriptors.

>> An alternative to default overriding, is to specify explicit 
>> overriding rules. But also in this case we have to check 
>> whether two descriptors are in conflict or not, based on 
>> their definition.
> 
> I am not sure we need that, because in the end we only want to provide
> information.  
> I don't think we should make decisions on what is desired and what not.

I agree. But, as I said above, one thing is making decisions about 
contradictory statements made by different people, but how can I decide 
about contradictory statements made by the same person? Should we 
consider such scenarios as equivalent?

An objection to this may be that in the real world this may frequently 
happen. Actually, DRs released by an organisation are not created by the 
organisation itself, but by people working for it, who can have 
different opinions. Maybe we can address this issue by using <maker> for 
the person who actually created the DR, and <certifiedBy> (<endorser>?) 
for the organisation releasing the DR. But, again, what if the same 
person of the same organisation makes contradictory statements?

Andrea

Received on Thursday, 14 February 2008 15:00:29 UTC