Re: [TF-PP] Duplicates and simple property path expressions.

On 2010-3-5 11:56 , Steve Harris wrote:
> On 5 Mar 2010, at 08:02, Ivan Herman wrote:
>>
>> On 2010-3-4 15:53 , Andy Seaborne wrote:
>>> In
>>> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2010Feb/0018.html
>>>
>>>
>>>
>>> Rob Vesse observes that "no duplicates" is at odds with the expansion of
>>> a simple path (a term we have already decided to remove) into a triple
>>> pattern because there is an implicit projection going on.
>>>
>>>  {?x foaf:knows{2} ?y}
>>> is not quite the same as
>>>  { ?x foaf:knows ?z . ?z foaf:knows ?y }
>>> when there are acylic components/
>>> e.g.:
>>>
>>> :a foaf:knows :b .
>>> :a foaf:knows :c .
>>> :b foaf:knows :d .
>>> :c foaf:knows :d .
>>>
>>> because ?z is projected away.
>>>
>>> We can specify either way:
>>>
>>> + emphasis that certain property paths are the same as the triple
>>> expansion form, and not have the striuct no duplicates rule (this
>>> reintroduces the simple property paths concept) or
>>>
>>> + note, and provide an example, that they are not exactly equivalent.
>>>
>>> I prefer the latter - keep the "no duplicates" situation.
>>>
>>
>> So do I. As a user, who does not know about the algebra and such, having
>> ?x/:a and ?y/:d to the {?x foaf:knows{2} ?y} seems like the natural
>> answer...
> 
> Actually, naively I'd expect {?x foaf:knows{2} ?y} to do the same as {?x
> foaf:knows ?tmp . ?tmp foaf:knows ?y}. On the other hand, I suppose
> people might not care about cardinality in the property path case? Not
> strong feelings I guess.
> 
> What does this do to list handling though? If you have
>    :x :p (1 2 2 3) .
> and
>    { :x rdf:rest*/rdf:first ?y } [or whatever the right combo is]
> won't you get
>    ?y = 1, 2, 3
> as a result?
> 

Ouch. You got me there:-(

Ivan


> - Steve
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF   : http://www.ivan-herman.net/foaf.rdf
vCard  : http://www.ivan-herman.net/HermanIvan.vcf