- From: Michael Kifer <kifer@cs.stonybrook.edu>
- Date: Sun, 25 Jul 2010 22:30:58 -0400
- To: Sandro Hawke <sandro@w3.org>
- CC: public-rif-wg <public-rif-wg@w3.org>
OK, I think I am getting closer. You should have started with this. So, it is not a matter of inference and its completeness but rather of which rule exactly a particular RDF graph encodes, right? That is, if we do not use lists explicitly in RDF then there is no way to encode a&b&c=>d correctly? That it will be instead be encoding a set of rules that in addition includes a&b=>d, a&c=>d, a=>d, etc. This looks like a strong argument. I haven't thought about it. Dave, Harold, do you see a way around this problem? michael On Sun, 25 Jul 2010 22:10:52 -0400 Sandro Hawke <sandro@w3.org> wrote: > I think we're misunderstanding each other. It's possible I'm just > wrong (as you seem to think), but I don't think so. I can try to > refine my language, but it's not clear to me how to do that right now. > > So let me try a completely different line of argument for a minute. > Please try to read between the lines of my poor command of the language > of philosophy and formal logic: > > 1. RDF is a KR. An RDF graph is a logical statement; it makes claims > about some world. > > 2. If someone makes a claim by stating RDF Graph G, they are implying > all the claims made by all the graphs entailed by G. > > 3. While there may be some debate about which logics (and therefore > which entailments) are appropriate/standard, RDF Simple Entailment > certainly is. So whenever you say something as an RDF Graph G, you are > implying all the claims of all the subgraphs of G. > > 4. Assume we have the RIF rule, a&b&c=>d, and it's encoded in graph G > as node R. This means the agent stating G is claiming that R says > a&b&c=>d. > > 5. If we use the repeated-properties mapping for that step-4 encoding, > then there will be subgraphs of G which describe R as saying a&b=>d, > a&c=>d, a=>d, etc. > > 6. If the agent stating G is also asserting R, then (by #5) the agent > is also implying a&b=>d, a&c=>d, c=>d, etc. If c is true, the agent > will have implied d. Consumers trusting that agent may justifiably > infer d. If it turns out a or b was false, this conclusion is wrong. > So, this is bad. > > 7. With the list-style mapping, the only subgraphs of G in which R > encodes a RIF Document are those in which that encoded document has the > same RIF meaning. This is because the list-style mapping is essentially > fragile; all the non-trivial subgraphs simply don't describe a RIF > document. So we avoid #6 badness. > > I'm sorry I don't know how to say this in the proper mathematical > language. I hope it still makes sense. Let me know which steps I need > to expand on/clarify. > > -- Sandro
Received on Monday, 26 July 2010 02:31:32 UTC