- From: Golda Velez <gv@btucson.com>
- Date: Mon, 4 Feb 2008 00:38:17 -0700
- To: editor@content-wire.com
- Cc: semantic-web@w3.org
Hi Paola I think, that RDF is capable of expressing complex clauses by making a complete clause into the object (or subject) of an assertion. By using a 'blank node' you can make pretty complex statements that would definitely annoy your English prof. But here I'm out of my depth and I'll defer to others. I do agree, it can get messy, and I'm interested to see what others have to say on this subject re best practices for representing complex statements... Hm - unfortunately the Best Practices working Group is closed http://www.w3.org/2001/sw/BestPractices/ - is it really all worked out already? The closest doc I can find is http://www.w3.org/TR/swbp-n-aryRelations/ but that doesn't quite fill the bill here... --Golda On Sunday 03 February 2008 9:56, editor@content-wire.com wrote: > Golda > > I am still learning about RDF, > Let me give you my take, so that others may correct me. > > Based on what I read so far, I believe that RDF is not designed to support > complex statements of natural language, although I am sure ways around can > be found, albeit 'not elegant' > > Most natural languages have developed to support sophisticated reasoning and > logic > They have done by allowing for two classes 'simple clauses' (subject > predicate object), which can stand alone, and can be supported by RDF > nicely, and 'complex clause's, which are formalisms to express more > structured and less linear logical relationships, they depend on a simple > clause to make sense, like the examples that you give (the second part of > your sentence is a relative clause, either coordinate or subordinate > see the diagram below > > http://papyr.com/hypertextbooks/grammar/cmplx1.htm > > > > > > > <Which> leads to is a subordinate, relative clause, meaning: cannot stand on > its own, > it needs another simple sentence to hang from. > Such structure is not supported by RDF > > So, If I were obliged to use RDF to do this, I would use > > - with a chain of triples, and assign arbitrary values to each triple > <which> could be a predicate, for example in current RDF syntax, because > predicates can be > 'absolutely antying'. From a design and natural language perspective, IMHO > this is not robust nor elegant nor efficient > or > allow RDF to express n'tuple (not just three values statements) assigning as > many predicates and as many objects as your sentence statement requires - > equally messy. > > My suggestion would be: augment and evolve RDF syntax to support > subordinate clauses more complex grammatical structures as illustarted above > This can be easily and quickly done, and would solve the problem > we just need to update the RDF spec a bit > > If I have to use the grammatical structure of a simple sentence to expresse > complex sentences, is like > trying to express modern literature using the language spoken by Neanderthal > Nothing wrong with it, but not adequate to the evolutionary stage > > Please let me know where I am wrong > > cheers > > Paola Di Maio > > > > > ------------------ > > How do I find the right vocabulary to express this statement? And for > > others > > to agree and disagree with it, what is the best way to give it its own URI > > or > > other identifier so other statements can be made about this statement? > > > > I think that in the mucky real world, making and responding to statements > > of > > this type of complexity would be useful. Should we just stick to English > > or > > our other native tongue? > > > > --Golda > > > >
Received on Monday, 4 February 2008 07:34:40 UTC