Re: propose to make repeated-properties additive

On 9/22/2015 15:03, Karen Coyle wrote:
>
> Holger, whether they can "afford it" or not isn't relevant, I don't 
> believe. The fact is that there is quite a lot of use of schema.org. 
> I'm told that it is now used on about 1/3 of all web sites. 

Of which 99.9% is likely auto-generated by platforms such as Wordpress, 
without web developers even knowing it :)

> The library world's main database, http://worldcat.org, has schema.org 
> encoding for a major percentage of its over 300 million items, and 
> that's just one database. And, as I say repeatedly, Dublin Core 1.1 is 
> one of the most used vocabularies in LoD, and it does not define 
> ranges for any of its properties, and usage (literal or IRI) varies.

Ok, then you can hopefully provide examples of *constraints* on such data.

Anyway, I have moved beyond that and have acknowledged that such use 
cases do exist. I have in fact implemented a solution to that issue.

>
> I'm concerned that it will be difficult to explain the need to use the 
> "sh:qualified*", and what the implications are of its use. I'm also 
> concerned that readers of the spec will miss the distinction. My goal, 
> therefore, is to clarify the use of "qualified" for users.

Yes, same here. We need to clarify this distinction.

>
>>
>> Anyway, let's look at what it would take to express such things. I
>> believe we need to continue to keep general property constraints and
>> qualified property constraints separate, because the majority of
>> constraints is not qualified but applies to all values.
>
> I'm not at all sure that this is true (it's a gut feeling unless 
> someone can produce some actual data), 

Others may have better data, but it is pretty obvious that QCRs in OWL 
are barely used compared to normal cardinality restrictions. Take a look 
at http://stats.lod2.eu/properties?search=cardinality or look at the 
number of google hits for owl:minCardinality (11,200) vs 
owl:minQualifiedCardinality (2,850)

> and I don't see this as a good reason to keep them separate. There may 
> well be other reasons, but to me the key is that it be easy to 
> understand -- for those people who aren't already fully immersed in OWL.

Yes, and the current SHACL design is IMHO intuitive for anyone who never 
had any exposure to OWL or RDF, e.g. from an object-oriented world. When 
you see a minimum cardinality in UML it aligns exactly with SHACL.

>
>
> In order to
>> express your use case above, we would need to extend the
>> sh:qualifiedValueShape mechanism so that it can also work with literals.
>> Currently my implicit assumption was that the focus node cannot be a
>> literal, but this is probably an unnecessary restriction. I have just
>> opened a ticket for that micro decision.
>
> I don't know what I said that would indicate that a shape have a 
> literal as subject -- is that what you mean here? That sounds like 
> something that violates the basis of RDF, so I don't know why it would 
> be needed. Did I misunderstand what you meant?

Yes.

Holger

Received on Tuesday, 22 September 2015 05:47:11 UTC