Re: ISSUE-137: Re: rdf language tag contraints

On 19/03/2016 2:06, Karen Coyle wrote:
> I can think of these use cases:
>
> 1. literal must have a language tag - any language tag

This can already be expressed, e.g. sh:datatype rdf:langString.

> 2. literal must have a specific language tag
> 3. literal must have one of a list of language tags

We could add sh:langShape

ex:MyShape
     a sh:Shape ;
     sh:property [
         sh:predicate rdfs:label ;
         sh:langShape [ sh:in ( "de" "fr" "en" ) ] ;
     ] .

which would also allow any other condition on the language tag (if 
needed). It would use sh:hasShape internally. Alas sh:in may not be 
ideal as it would not do the correct langMatches semantics.

>
> The more difficult one is:
>
> 4. there must be n properties, each with one of a list of language 
> tags, no repeats, e.g. you must have 3 SKOS preflables, one each @en 
> @sp @fr

I think that's just a combination of existing features, e.g.

ex:MyShape
     a sh:Shape ;
     sh:property [
         sh:predicate skos:prefLabel ;
         sh:minCount 3 ;
         sh:maxCount 3 ;
         sh:langShape [ sh:in ( "de" "fr" "en" ) ] ;
         # or: sh:uniqueLang true  instead of minCount/maxCount
     ] .

Any variation of sh:langShape can be added to the Core. The only 
decision we need to make is whether it's common enough and whether we 
have the resources to give justice to this topic (any new feature 
requires tests, documentation, discussions).

As long as we cannot even agree on the basic syntax of SHACL I would 
postpone everything else.

Holger

>
> There may be others, that's what's off the top of my head.
> kc
>
> On 3/18/16 1:09 AM, Bart van Leeuwen wrote:
>> Hi Holger,
>>
>> The requirement itself has already been discussed in the UCR document.
>>
>> The background for my interest is using shapes for UI generation.
>> If you generate a UI from RDF data it makes sense to have the litterals
>> displayed in the preferred user language, or to be able to indicate this
>> is not available in the data.
>> When you have a shape describing the UI constraints you can instantly
>> find out if the data to be displayed is actually available in the
>> language you need.
>>
>> So if I would use e.g. sh:language or sh:languageIn on a property it
>> would be valid if at least one of the property values matches the 
>> language.
>>
>> As much as I appreciate the power of the shacl extensions as described
>> in 6.1 I think that the language tag is such fundamental part of the RDF
>> spec that it should be in core.
>> We also have a SHACL core constraint for datatypes which are also part
>> of the RDF spec
>>
>> Met Vriendelijke Groet / With Kind Regards
>> Bart van Leeuwen
>>
>> ##############################################################
>> # twitter: @semanticfire
>> # netage.nl
>> # http://netage.nl <http://netage.nl/>
>> # Esdoornstraat 3
>> # 3461ER Linschoten
>> # tel. +31(0)6-53182997
>> ##############################################################
>>
>>
>>
>> From: Holger Knublauch <holger@topquadrant.com>
>> To: public-data-shapes-wg@w3.org
>> Date: 18-03-2016 03:16
>> Subject: ISSUE-137: Re: rdf language tag contraints
>> ------------------------------------------------------------------------
>>
>>
>>
>> Hi Bart,
>>
>> could you clarify your requirement:
>>
>> 1) Do you want to express that at least one of the values of a given
>> property has a certain language tag?
>>
>> 2) Do you want to express that all values of a given property have a
>> certain language tag?
>>
>> On the question of usability, I assume you know that the example in 6.1
>> is not how users would typically encounter this. SHACL has the extension
>> mechanism that allows new high-level terms to be defined, even if they
>> missed the boat of the current SHACL Core. It would then look like
>>
>> ex:MyShape
>>      a sh:Shape ;
>>      sh:property [
>>          sh:predicate rdfs:label ;
>>          lang:hasValueWithLang "en" ;
>>      ] .
>>
>> assuming that someone has published a corresponding constraint component
>> as a library with the namespace prefix "lang:" (which would be quite
>> trivial to do and would exploit the power of the web instead of
>> delegating everything to a committee).
>>
>> Thanks,
>> Holger
>>
>>
>> On 15/03/2016 7:51, Bart van Leeuwen wrote:
>> Karen,
>>
>> I created ISSUE-137, mentioning ISSUE-35
>> I guess this is the proper way to do it.
>>
>> Met Vriendelijke Groet / With Kind Regards
>> Bart van Leeuwen
>>
>> ##############################################################
>> # twitter: @semanticfire
>> # netage.nl
>> # _http://netage.nl_ <http://netage.nl/>
>> # Esdoornstraat 3
>> # 3461ER Linschoten
>> # tel. +31(0)6-53182997
>> ##############################################################
>>
>>
>>
>> From: Karen Coyle _<kcoyle@kcoyle.net>_ <mailto:kcoyle@kcoyle.net>
>> To: Bart van Leeuwen _<bart_van_leeuwen@netage.nl>_
>> <mailto:bart_van_leeuwen@netage.nl>
>> Cc: _public-data-shapes-wg@w3.org_
>> <mailto:public-data-shapes-wg@w3.org>, Arnaud Le Hors
>> _<lehors@us.ibm.com>_ <mailto:lehors@us.ibm.com>
>> Date: 14-03-2016 22:42
>> Subject: Re: rdf language tag contraints
>> ------------------------------------------------------------------------
>>
>>
>>
>> ISSUE-35 is closed. I think this needs to be a new issue, can refer to
>> ISSUE-35 outcome, and the message is: add this functionality to the
>> SHACL specification. (Arnaud, if this isn't the right procedure, pls
>> advise.)
>>
>> kc
>>
>> On 3/14/16 1:45 PM, Bart van Leeuwen wrote:
>>  > I'll pick it up,
>>  >
>>  > it is still related to ISSUE-35 right ?
>>  >
>>  > Met Vriendelijke Groet / With Kind Regards
>>  > Bart van Leeuwen
>>  >
>>  > ##############################################################
>>  > # twitter: @semanticfire
>>  > # netage.nl
>>  > # _http://netage.nl_ <http://netage.nl/><_http://netage.nl/_>
>>  > # Esdoornstraat 3
>>  > # 3461ER Linschoten
>>  > # tel. +31(0)6-53182997
>>  > ##############################################################
>>  >
>>  >
>>  >
>>  > From: Karen Coyle _<kcoyle@kcoyle.net>_ <mailto:kcoyle@kcoyle.net>
>>  > To: Bart van Leeuwen _<bart_van_leeuwen@netage.nl>_
>> <mailto:bart_van_leeuwen@netage.nl>
>>  > Cc: _public-data-shapes-wg@w3.org_ 
>> <mailto:public-data-shapes-wg@w3.org>
>>  > Date: 14-03-2016 21:43
>>  > Subject: Re: rdf language tag contraints
>>  > 
>> ------------------------------------------------------------------------
>>  >
>>  >
>>  >
>>  > I believe we need an issue to make sure this gets discussed. I can
>>  > create that if you wish, or you can do it.
>>  >
>>  > kc
>>  >
>>  > On 3/14/16 9:17 AM, Bart van Leeuwen wrote:
>>  >  > Hi Karen,
>>  >  >
>>  >  > Thanks for the follow up, does this mean that a constrained in 
>> SHACL
>>  >  > should also be added ?
>>  >  >
>>  >  > Met Vriendelijke Groet / With Kind Regards
>>  >  > Bart van Leeuwen
>>  >  >
>>  >  > ##############################################################
>>  >  > # twitter: @semanticfire
>>  >  > # netage.nl
>>  >  > # _http://netage.nl_
>> <http://netage.nl/><_http://netage.nl/_><_http://netage.nl/_>
>>  >  > # Esdoornstraat 3
>>  >  > # 3461ER Linschoten
>>  >  > # tel. +31(0)6-53182997
>>  >  > ##############################################################
>>  >  >
>>  >  >
>>  >  >
>>  >  > From: Karen Coyle _<kcoyle@kcoyle.net>_ <mailto:kcoyle@kcoyle.net>
>>  >  > To: _public-data-shapes-wg@w3.org_
>> <mailto:public-data-shapes-wg@w3.org>
>>  >  > Date: 14-03-2016 15:44
>>  >  > Subject: Re: rdf language tag contraints
>>  >  >
>> ------------------------------------------------------------------------
>>  >  >
>>  >  >
>>  >  >
>>  >  > Bart, thanks. I followed up on the history and it was indeed 
>> voted to
>>  >  > become a requirement, but wasn't added. I will add it to the
>>  >  > requirements list in the UC&R.
>>  >  >
>>  >  > kc
>>  >  >
>>  >  > On 3/13/16 9:46 AM, Bart van Leeuwen wrote:
>>  >  >  > Hi All,
>>  >  >  >
>>  >  >  > Seeing all the productive discussions and positive vibe on the
>> list I
>>  >  >  > sat down and read the current Editors draft of SHACL [1]
>>  >  >  >
>>  >  >  > One of the things that strike me is that I cannot have a simple
>>  >  >  > constraint on a language tag.
>>  >  >  >
>>  >  >  > e.g. I need rdfs:label with '@en' for my application.
>>  >  >  >
>>  >  >  > I do see a example in the "Native Constraints" section [2] 
>> but that
>>  >  >  > looks rather complex for my use case.
>>  >  >  > On the issue tracker I did find ISSUE-35 [3] which resolves in
>>  > allowing
>>  >  >  > constraints on the language tag.
>>  >  >  >
>>  >  >  > I tried searching the mailing list archive but 'Language' is so
>>  >  >  > ambiguous that it didn't help me answer this question from the
>>  > archive.
>>  >  >  >
>>  >  >  > Why is the use case mentioned in Issue 35 not there any 
>> more, and
>>  > why is
>>  >  >  > there no simple way constraint on the language of a literal?
>>  >  >  >
>>  >  >  > [1] _http://w3c.github.io/data-shapes/shacl/_
>>  >  >  > [2]
>>  > 
>> _http://w3c.github.io/data-shapes/shacl/#an-example-native-constraint_
>>  >  >  > [3] _https://www.w3.org/2014/data-shapes/track/issues/35_
>>  >  >  >
>>  >  >  > Met Vriendelijke Groet / With Kind Regards
>>  >  >  > Bart van Leeuwen
>>  >  >  >
>>  >  >  > ##############################################################
>>  >  >  > # twitter: @semanticfire
>>  >  >  > # netage.nl
>>  >  >  > # _http://netage.nl_ <http://netage.nl/>
>>  > <_http://netage.nl/_><_http://netage.nl/_><_http://netage.nl/_>
>>  >  >  > # Esdoornstraat 3
>>  >  >  > # 3461ER Linschoten
>>  >  >  > # tel. +31(0)6-53182997
>>  >  >  > ##############################################################
>>  >  >
>>  >  > --
>>  >  > Karen Coyle
>>  >  > _kcoyle@kcoyle.net_
>> <mailto:kcoyle@kcoyle.net><http://kcoyle.net/>_http://kcoyle.net_
>> <http://kcoyle.net/>
>>  > <_http://kcoyle.net/_><_http://kcoyle.net/_>
>>  >  > m: 1-510-435-8234
>>  >  > skype: kcoylenet/+1-510-984-3600
>>  >  >
>>  >  >
>>  >  >
>>  >
>>  > --
>>  > Karen Coyle
>>  > _kcoyle@kcoyle.net_ <mailto:kcoyle@kcoyle.net>_http://kcoyle.net_
>> <http://kcoyle.net/><_http://kcoyle.net/_>
>>  > m: 1-510-435-8234
>>  > skype: kcoylenet/+1-510-984-3600
>>  >
>>  >
>>  >
>>
>> -- 
>> Karen Coyle_
>> __kcoyle@kcoyle.net_ <mailto:kcoyle@kcoyle.net>_http://kcoyle.net_
>> <http://kcoyle.net/>
>> m: 1-510-435-8234
>> skype: kcoylenet/+1-510-984-3600
>>
>>
>>
>>
>>
>

Received on Saturday, 19 March 2016 01:02:04 UTC