Re: isolating shapes in named graphs

On 11/19/14, 9:14 PM, Eric Prud'hommeaux wrote:
>
> This is a critical architectural issue. If we weave mutually
> inconsistent data into the shapes architecture, we over-constrain its
> use in ways that will have long-term consequences.

The same argument can be applied in the reverse direction: if we 
separate shapes from classes, we over-complicate its use from the data 
models that exist in parallel. For the majority of use cases you would 
end up with Shape objects that are mirroring classes, while it would 
indeed be more consistent (and easier to look up from a Linked Data 
perspective) if they were as aligned as possible. In ShExC you 
furthermore end up with two different syntaxes for virtually the same 
entities - one is triple-based and the other based on regular expressions.

Having said this, please take a look at my proposal to merge stand-alone 
Shapes into SPIN structures, based on the introduction of just a single 
helper function, which spawns off recursive constraint checking engines. 
I believe this may lead to a solution that we can all live with, and 
that is still easy enough to explain and consistent to implement.

The rest then becomes syntactic sugar.

Holger

Received on Wednesday, 19 November 2014 12:37:29 UTC