Re: shapes distinct from classes docs

On 2/7/15 12:49 AM, Eric Prud'hommeaux wrote:
> I believe that SPARQL is about 10x the size and complexity of what one 
> needs to handle the stuff that's common to ShEx, Resource Shapes and 
> Description Set Profile. I hope that one can have a shapes engine 
> without having to have a SPARQL implementation.

Yes absolutely. While we require that SPARQL is made part of the 
official standard, LDOM also includes a notion of "profiles" - 
essentially a catalog of templates such as cardinality + valueType + 
valueShape. Many applications will chose to only support such light 
profiles and they can perfectly do so, similar to how OWL has different 
profiles. As I had outlined in my proposal on ending the controversies, 
it will be fine for ShEx-like engines to evaluate LDOM shape 
declarations without any SPARQL at all. IBM had similar requirements for 
their OSLC scenarios.

Holger

Received on Saturday, 7 February 2015 00:22:09 UTC