- From: Holger Knublauch <holger@topquadrant.com>
- Date: Sun, 08 Feb 2015 10:10:52 +1000
- To: public-data-shapes-wg@w3.org
This discussion reminds me of a top-down design in which we (the WG members) believe we know best what our users would want in the future. I doubt that either of us has enough foresight about that, just like previous WGs did not anticipate that most people would abuse rdfs:domain and range to work just like in OO systems. Even if we pretend that classes cannot serve as shapes, people will do this anyway, and tools will support this anyway, because it makes perfect sense for anyone who enters this world from mainstream technologies. Even if we pretend that LDOM is not a modeling language, for all practical purposes it can be regarded as a modeling language. If the WG members cannot agree on this situation, here is my proposal: - Let's not publish any Primer at all - Let's focus on the formal spec - Keep the formal spec short and precise, yet flexible - If needed, use neutral examples with URIs such as ex:Shape and ex:SubClass - Write test cases that implementers can use to verify the correctness It will be perfectly possible for 3rd parties to publish books and tutorials such as "Web-friendly Object Oriented Modeling with XY" and "Serendipitous reuse of linked data with XY Shapes", and the users pick the starting point they find most attractive. Just a few implementers will read the formal spec. As a nice side effect, we have one less deliverable to worry about. Holger On 2/8/15 1:05 AM, Eric Prud'hommeaux wrote: > > I'm skeptical that it's a common occurance in sensible modeling, but > I'm certainly happy to be shown otherwise. Its possible that our > disagreement stems from different starting conditions. Here are mine: > > Much of the value of RDF stems from "serendipitous reuse". > > The prominent examples should use the core shapes language. > > Physical laws like area aren't typical of business logic. > > >> peter >>
Received on Sunday, 8 February 2015 00:11:29 UTC