W3C home > Mailing lists > Public > public-data-shapes-wg@w3.org > March 2016

Re: Selected problems with Proposal 4

From: Holger Knublauch <holger@topquadrant.com>
Date: Fri, 11 Mar 2016 14:22:30 +1000
To: public-data-shapes-wg@w3.org
Message-ID: <56E24806.3090203@topquadrant.com>
On 11/03/2016 13:28, Karen Coyle wrote:
>
> On 3/10/16 5:01 PM, Holger Knublauch wrote:
>>> In my proposal this would be
>>>    ex:foo a sh:Shape ;
>>>     sh:property ( ex:guru [ sh:class ex:Person; sh:class ex:Preacher ]
>>> ) .
>>> The current syntax results in shapes that are harder to analyze by 
>>> tools.
>>
>> No, the tools will have to do even more work in your case, because there
>> are more syntax variations to express the same thing (in your approach,
>> multiple fillers may exist, but also sh:and and now multiple sh:class
>> directly).
>>
>> Furthermore, it is trivial to pick certain cases to support your case.
>> We need to look at the big picture, and many different examples. I would
>> consider the case of "Person and teacher" to be not very common, but
>> others will of course disagree and make the claim it's critical. So how
>> would anyone decide that? It will always be subjective.
>
> Is this construct specific only to sh:class? from the examples in 
> proposal 4 it seems to be more general, so perhaps more examples are 
> needed.

It's not specific to sh:class. The only real question here is whether we 
want special syntactic sugar for a case where a property such as 
sh:class shows up twice for the same context. The price of this 
syntactic sugar is that no constraint type can take more than one 
parameter (such as sh:pattern/sh:flags) because otherwise it would not 
be clear which combination of values belong together. I do not believe 
this price is worth paying, as the case is not common, and easy 
work-arounds already exist.

Holger
Received on Friday, 11 March 2016 04:23:09 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:30:30 UTC