W3C home > Mailing lists > Public > public-rdf-shapes@w3.org > September 2016

Re: [CODE4LIB] SHACL Core abstract syntax

From: Stuart A. Yeates <syeates@gmail.com>
Date: Wed, 7 Sep 2016 20:32:04 +1200
Message-ID: <CAC_Lu0YeNO+LnVS3daRTmUNvz2dNCPkyjr0wU0H-Stafgvcb1A@mail.gmail.com>
To: Karen Coyle <kcoyle@kcoyle.net>
Cc: public-rdf-shapes@w3.org
On Wed, Sep 7, 2016 at 8:26 AM, Karen Coyle <kcoyle@kcoyle.net> wrote:

> Thanks, Stuart. Some replies below:
>
> On 9/6/16 3:42 AM, Stuart A. Yeates wrote:
>
>> This is fabulous.
>>
>> Some feedback, based on a relatively quick look (I may have overlooked
>> things)
>>
>> * In the examples, please use .example.org <http://example.org> (or
>> similar) rather than .example. More people will find it obvious.
>>
>
> It's true that .example is less "usual", but it is in the same RFC as
> example.com|org|etc.[1] We used it when trying to show different IRIs,
> and where "a.example.com" and "b.example.com" would be from the same
> domain. That said, we take your point and will consider other ways to make
> the different IRIs show up better.


If demonstrating IRIs is your goal, bite the bullet and use one with
non-ASCII characters.


> * It's not clear whether SHACL is checking against the RDF graph with or
>> without the implicit reverse relationships.
>>
>
> One thing that we didn't say here, and perhaps need to (although in a
> sense it belongs in the main document), is that SHACL, like SPARQL,
> operates over an RDF graph and does not modify it. Only the *explicit* data
> graph is in play. We think that we can make a brief statement in the
> introduction, since it obviously is a point of confusion.


Great.

So include an example of how to overcome this, killing both the ambiguity
and any objection in a single blow.


> * It would be good to have an example of the form:
>>
>> <user1> ex:relationship <user2>.
>>
>> and testing the shape against both <user1> and <user2>
>>
>
> We're thinking about how to do that. Do you have a specific example in
> mind? Otherwise, we'll probably include one that uses foaf:Person, which
> seems like an obvious choice.


The more I look at this, the more the foaf: namespace is a semantic web
in-joke. If the target audience is not those already immersed in the
semantic web, switch to something people have seen used for real problems,
like dc.

cheers
stuart

--
...let us be heard from red core to black sky
Received on Wednesday, 7 September 2016 08:32:35 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:02:44 UTC