- From: Rama Akkiraju <akkiraju@us.ibm.com>
- Date: Mon, 20 Aug 2007 22:33:35 -0400
- To: Jacek Kopecky <jacek.kopecky@deri.org>
- Cc: SAWSDL public list <public-ws-semann@w3.org>, public-ws-semann-request@w3.org
Jack, In my opinion, we should not drop section 3.7. If you recall, it was added to address a point that was repeatedly raised by some parts of the community that they would get most value from the spec if some guidance is given on how to represent preconditions and effects using SAWSDL spec. We did not want to make any explicit statements about preconditions and effects in the main spec. So, we chose to add some examples in the user guide document. To the extent possible, we should fix up the rules and keep it in the user guide. If the rules can't be fixed up to the point of satisfaction and verification, I suggest we move the section to appendix, with some caveats but not delete it completely. Regarding your point on coming up with more plausible conditions and effects, that would be good if anyone has any inputs. But I have already sent out emails to the public semantic interest group seeking examples on preconditions and effects at the time and have not received any response from anyone. The only piece of input we got is from Thomas (thanks to him for that) and we have incorporated those. Frankly, although the current conditions and effects are lame, they fit in the purchase order domain we have described throughout the document. More plausible conditions may require a lot of context and explanation of another example, which we have purposefully avoided to keep the examples simple. From that point of view, I think the current conditions are ok if not the best. Regards Rama Akkiraju Jacek Kopecky <jacek.kopecky@de ri.org> To Sent by: SAWSDL public list public-ws-semann- <public-ws-semann@w3.org> request@w3.org cc Subject 08/20/2007 03:00 usage guide section 3.7 is broken PM Dear all, I have reviewed the Usage Guide section 3.7 and found that, basically, the rule listings seem all to be broken to varying degrees (in fact, all of them say "syntax subject to verification" which should not be the case so close to finishing). These are just sample errors from those I have spotted: * in 3.7-1, swrlb:subtract tests for equality, i.e. that the delivery date and order date are exactly two days apart, not at least two days * the listings use many undefined things - in 3.7-1 the term shipment(), in 3.7-2 the symbols waitingTime, deliveryTime etc. (not present in 3.7-1, while the two listings should be equivalent) * 3.7-3 has no XML root element so it's not well-formed, nor does it define any namespaces * 3.7-4 references wsml#subtract-dateTimes-yielding-day which is undefined * the rule names, well hidden in 3.7-3 and different from 3.7-4, are not at all intuitive The document should be finished within a few days, and I see three options how we can end up with a consistent document in such a timeframe: 1. we drop all the concrete syntax listings of the rules (3.7-1 - 3.7-6), keep only the verbal description at the beginning of section 3.7 and update the model references in 3.7-8 to point to imaginary embodiments of rules, with descriptive names; 2. we drop the whole section 3.7, seeing how we can't really make it concrete, and it puts annotations in places undefined in our spec (wsdl:input and wsdl:output) anyway; 3. before the telcon, the editors (or anyone else interested in this) prepare listings (at least one for condition and one for effect) that are much more plausible than the current ones. Also note that the text immediately preceding 3.7 should be moved to the end of 3.7 if we decide to keep the section; it serves as a segue to section 4. Please comment; a course of action will be decided at the telcon tomorrow. 8-) Best regards, Jacek
Received on Tuesday, 21 August 2007 02:33:49 UTC