- From: Holger Knublauch <holger@topquadrant.com>
- Date: Thu, 27 Nov 2014 09:41:40 +1000
- To: public-hydra@w3.org
On 11/27/2014 1:29, tomasz@t-code.pl wrote: > Hi Holger > > I've also been thinking about using SPIN on the client. In addition to constraint validation I was also thinking about using spin:constructor to allow the client to initialize a state of a new resource. Even better the server can hand out "personalized" CONSTRUCTS to individual clients. Yes, I see very few use cases of spin:constructor in its current form - I guess that's because it cannot take parameters. It would be easy to attach a list of parameterized constructors to a class, e.g. as spin:ConstructTemplates. I'd certainly appreciate requirements and use cases so that we can add this in. Maybe a property such as spin:factory would be enough, that points from an rdfs:Class to spin:ConstructTemplate, and ?this would identify the new instance. > I think SPIN on clients can be a best so far realization Roy Fielding's Code-on-demand REST contraint. > > But please correct me if I'm wrong, AFAIK executing SPARQL is necessary but it's not enough to fully appreciate SPIN. And it's a shame that after all those years TopBraid's is probably the sole SPIN implementation. At least I haven't seen any other yet. Let's hope that the Shapes group is a stepping stone to producing something "similar to" SPIN with a W3C logo on it. If that happens, then I am confident that more implementations will emerge soon. From the six database vendors that I had contacted, all were very interested in this topic. And then there could be client side engines, a Sesame API etc. Plenty of opportunities for useful open source and thesis projects. The language is a relatively thin layer on top of SPARQL and therefore quite easy to implement. Thanks, Holger
Received on Wednesday, 26 November 2014 23:44:27 UTC