- From: Chris Welty <cawelty@gmail.com>
- Date: Mon, 23 Jul 2007 17:13:44 -0700
- To: "Boley, Harold" <Harold.Boley@nrc-cnrc.gc.ca>
- CC: "Public-Rif-Wg (E-mail)" <public-rif-wg@w3.org>
Thanks, Harold - some further clarifications below - Chris Boley, Harold wrote: > Chris, > > Thanks for your notes. Michael changed a lot already and will respond > (perhaps tomorrow, because of CEST timezone). I worked on 2.1.1.3. > > -- Harold > > > Note: The section numbers Chris is referring to are generated by > Sandro's wiki-tr, as used at http://www.w3.org/2005/rules/wg/wiki/Core. > While the "7/20 BLD" has changed meanwhile, you can always get an > up-to-date version there. > > > 2.1.1.3 > > >> It would be best if the correspondence between the ASN and the ebnf >> were more obvious. Perhaps this will be generated automatically at >> some point, but for now the productions (nonterminals) should have the > >> same names as the classes in the ASN. I'm happy changing the names of > >> the ASN classes to CONJUNCTION, etc. > > The usual names And/Or are used as visible ASN class names, > as XML element names, and as EBNF prefix operator names. > For the EBNF (only) there is a need to distinguish between > a terminal name like 'And' and an invisible nonterminal name > like CONJUNCTION that generates an entire expression applying > the operator to arguments as in 'And' '(' CONDITION* ')'. Yes - my point is that the nonterminals should have the same names as the ASN classes, because I suspect that is what any automatic translation would generate - you see what I mean? Deciding which terminal strings (e.g. "And" but also including things like commas, parens, etc) need to be in the concrete ebnf syntax is another issue. >> * It seems to me a cleaner syntax is to remove existentials from >> conditions and unify all quantifiers outside the rule (as with >> universal now), and add a restriction in horn that existentially >> quantified vars cannot appear in the conclusion. > > The RIF Basic Condition Language is meant to be reusable also > (in PR and) outside the context of any rule, stand-alone (e.g. > for queries and integrity checking), where existentials cannot > be rewritten as universals in the Example 3b.->3a. manner. Oh, you misunderstood - I did not mean eliminate existentials from the syntax entirely, I meant move them outside conditions into some sort of wrapper or context (like, rules). This would make, at least, the rule syntax cleaner. -- Dr. Christopher A. Welty IBM Watson Research Center +1.914.784.7055 19 Skyline Dr. cawelty@gmail.com Hawthorne, NY 10532 http://www.research.ibm.com/people/w/welty
Received on Tuesday, 24 July 2007 00:14:20 UTC