Tue Feb 4 14:15:02 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu JOIN :#joint-committee Tue Feb 4 14:15:08 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :logger, pointer? Tue Feb 4 15:06:41 2003: :mdean!-mdean@216.229.183.232 JOIN :#joint-committee Tue Feb 4 15:07:06 2003: :mdean!-mdean@216.229.183.232 PRIVMSG #joint-committee :hi sandro - thanks for joining - are you willing to scribe today? Tue Feb 4 15:25:04 2003: :mdean!-mdean@216.229.183.232 QUIT :Ping timeout Tue Feb 4 15:41:55 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net JOIN :#joint-committee Tue Feb 4 15:59:08 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Oh, sure. Heh. I'll try at least. Tue Feb 4 16:02:44 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Present: Mike, Andy, .... Tue Feb 4 16:02:51 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :... Stephan Tue Feb 4 16:02:54 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :... Ian Tue Feb 4 16:03:01 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :... Deb Tue Feb 4 16:03:04 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :... Sandro Tue Feb 4 16:03:12 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :... Ora Tue Feb 4 16:04:45 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: the first draft of requirements doc is postponed to next week; in general "2 weeks" is a good period between drafts Tue Feb 4 16:04:58 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :+Pat Tue Feb 4 16:05:38 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :mike: We set a slightly too agressive target for getting drafts out; I sent some ideas to the list this morning Tue Feb 4 16:05:55 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Stefan adapted the OWL requirements as a template for us to use Tue Feb 4 16:06:05 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: I also sent something out like 2 minutes ago Tue Feb 4 16:06:18 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :+Benjamin Tue Feb 4 16:06:30 2003: :dlm!+dlm@deborah.Stanford.EDU JOIN :#joint-committee Tue Feb 4 16:07:02 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Message was a "configurator" example; pointer to IJCAI example Tue Feb 4 16:07:24 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: we had a series of configurators at AT&T (17 at last count), using DL and Rules Tue Feb 4 16:07:50 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: The telecom domains are kind of hard to read, so the IJCAI paper is about a home-stereo configurator Tue Feb 4 16:08:21 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: OWL is more expressive than CLASSIC, so some examples are different (rules no longer needed) Tue Feb 4 16:08:44 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Computational example: when you have a value somewhere, you can fire a rule to do some calculations and add more data. Tue Feb 4 16:09:06 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Also a rule to avoid lots of named classes. Tue Feb 4 16:09:23 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: (those are the 3 kinds of rules usage in our configurator work.) Tue Feb 4 16:10:35 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: AT&T research web site cleared some stuff -- some of my old stuff is missing. There should be a redirection at some point. Tue Feb 4 16:11:33 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: It's important to distinguish between uses of rules which make up for weaknesses in DL (eg Classic), AND the people using it as a production-rule system where the rules had acccess to DL Tue Feb 4 16:12:12 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: RIght. Yes, most uses are no longer needed because of OWL's greater power. Tue Feb 4 16:12:31 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Benjamin: Were rules used because they were more convenient? Tue Feb 4 16:12:41 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Mostly just CLASSIC wasn't powerful enough. Tue Feb 4 16:13:10 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Filter example: you could say it in Classic, but it would clutter up your KB. Tue Feb 4 16:14:38 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: (more explanation of filters/rules in Classic) Tue Feb 4 16:15:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: (prompted by Pat) Tue Feb 4 16:15:36 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: Obviously you don;t need these in OWL, because you can say complex-desc IMPLIES complex-desc, which ytou couldnt say in CLASSIC. Tue Feb 4 16:15:56 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: that depends what you allow in the rule head, or has that been decided? Tue Feb 4 16:16:05 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: Tue Feb 4 16:16:20 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Benjamin: Is the clutter issue a performance one? Tue Feb 4 16:16:49 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: It's more human-cognative load. We also had a meta-language to allow classes to be hidden etc. Tue Feb 4 16:17:19 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: This is back to CLASSIC's you can only attach rules to named-classes. Tue Feb 4 16:17:54 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: ALso useful for debugging -- look at rules associated with a class [?] Tue Feb 4 16:18:35 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Benjamin: Maybe you could have used some of the available mechanisms? "UsuallySuppressClass", foo is a subclass of that, etc. Tue Feb 4 16:19:02 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: But then you want visible subclasses of foos, and you can't Tue Feb 4 16:19:12 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: THis is why you need classes of classes, guys. Tue Feb 4 16:19:33 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: [?] Tue Feb 4 16:19:46 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :?: something about naming things Tue Feb 4 16:20:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: ...? Tue Feb 4 16:20:08 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Why are you recommending this Deb? Tue Feb 4 16:20:44 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :Mike: want to track potential requirements/objectives/non-objectives during our discussions Tue Feb 4 16:20:53 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: having a lot of unnamed stuff in XML is a pain Tue Feb 4 16:21:35 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: It's cognative overload stuff. You can solve this with smarter interfaces, but it's often more work. Tue Feb 4 16:21:36 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :+? Tue Feb 4 16:22:25 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: WHen you can have conjunctive & partial-instantiated rule antecedents, you have a lot more (something) than just attaching them to classes. Tue Feb 4 16:22:33 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: Fine, that's what you want and what you have with OWL. Tue Feb 4 16:23:10 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Having rules associated with unnamed classes seems unnecessary and burdensome. Tue Feb 4 16:24:05 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: (repeat) need to distinguish between when it's part of a better DL, vs ...... + aggregation issssues, etc, not covered by OWL at all Tue Feb 4 16:24:09 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :+? Tue Feb 4 16:24:29 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :(someone in traffic now muted) Tue Feb 4 16:24:32 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :+Harold Tue Feb 4 16:24:43 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: DO you mean like lists? Tue Feb 4 16:25:10 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: Calculate total cost of peice of equipment being assembled; Rules were used for that.... Tue Feb 4 16:25:20 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: in stereo configurator. Tue Feb 4 16:25:36 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Monotonicity issues there? Tue Feb 4 16:25:52 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: We forced you to write the dependencies, and recalculate Tue Feb 4 16:25:59 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: aggregation is important. Tue Feb 4 16:26:39 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: Let's not overload same mechanism too much. Querying, Proceedural/Action, [Aggregation], ... Tue Feb 4 16:26:48 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: Why not? Look at SQL! Tue Feb 4 16:27:06 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: WHy not? FOr Interoperability. Two users with differnet styles and meanings Tue Feb 4 16:27:29 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: SQL's not a great example -- basic parts may be standard, but the fancy stuff is not. Tue Feb 4 16:27:40 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: But that's BECAUSE the core SQL was too small. Tue Feb 4 16:28:13 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Benjamin: Then there was a phase of dignification (cleanup) of that stuff in SQL, theory of rule systems for aggregation, etc. Tue Feb 4 16:28:53 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: Yes -- how should the rule language relate to the primitives of SQL? This could be an interesting comparison. (but recent SQL spec is about 900 pages!) Tue Feb 4 16:29:48 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Benjamin: aggregates,totals,counts, ... comparators (<, >), .... will need to be built in, in practice. We need some kind of proceedural attachements for these (whether you think of them as predicates or fiunctions). Tue Feb 4 16:30:30 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Logic programs have Aggregrate operator -- all the bindings from one atomic subgoal -- useing that as the basis of some other goal. Tue Feb 4 16:30:46 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee ::... procedural attachement to predicate/fuinction, with its own binding function. Tue Feb 4 16:31:12 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: with these two LP abstractgions, we can go a long way. Some SQL stuff is control-oriented and best avoided. Tue Feb 4 16:31:33 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Meta-requirement -- is this something that's important across several families of rules, not just one (like SQL). Tue Feb 4 16:31:43 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :[ Hear hear ] Tue Feb 4 16:31:59 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: What kind of builtins did you use in cwm? Tue Feb 4 16:32:16 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Not much -- I dont know about them much. Tue Feb 4 16:32:21 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: RuleML? Tue Feb 4 16:32:47 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: RuleML has some extensions, SituatedLP and EvaluatbleFunctions (nano) Tue Feb 4 16:33:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: We want modular builtins Tue Feb 4 16:33:13 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: EG the ISO Prolog builtings Tue Feb 4 16:33:57 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: declare relation with Pure Belief symbols. Things in the logic language. Not procedural. Tue Feb 4 16:34:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: Huh? Tue Feb 4 16:34:11 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Predicate shouldSendAnEmailTo Tue Feb 4 16:34:31 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: now you can reason about that, and we have a Java Class SendIt Tue Feb 4 16:35:03 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Make a distinction between a predicate (as in OWL/LP) and something which is a name which can be invoked as a procedure. Tue Feb 4 16:35:49 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: So if you want to do arithmetic, the convenient thing is to have a standard library of plugins with the names of a whole lot of methods (addition, ...) Tue Feb 4 16:36:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: What do they do if given an unbound variable? Tue Feb 4 16:36:14 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: it's important to be able to specify and constrain the binding patterns Tue Feb 4 16:36:59 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :..: SweetJess (RuleML -> Jess) translator discovered that Jess required all variables be unbound. Some allow some to be unbound; some allow even more unbound, doing backward-chaining query. Tue Feb 4 16:37:23 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: so it varies what kind of binding patterns you want to allow -- you need to be able to specify it all. Tue Feb 4 16:37:57 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: Called "Modes" and "Mode analysis" -- you could have multiple builtins Tue Feb 4 16:38:12 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: maluszinsky Tue Feb 4 16:38:25 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: On what basis is this called "clean"? Tue Feb 4 16:38:37 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Cyc and Snark are full of this kind of stuff Tue Feb 4 16:39:38 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: These systems tend to be formalized to either forward or backward, but I'm trying to clean 'em up. Guaranteeing completeness and soundness are interesting issues. Tue Feb 4 16:40:06 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: THis was a sensitive issue in CL. Some extensions we discussed broke other people's optimization machinery. Tue Feb 4 16:40:33 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: I'm writing a paper on this.... I'll send it along when I can. Tue Feb 4 16:40:51 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :logger, pointer? Tue Feb 4 16:41:03 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stephan: Next week, use case document will be ready Tue Feb 4 16:41:17 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: I suggest someone else work on Features document Tue Feb 4 16:41:37 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: I'd be interesting on helping put that together Tue Feb 4 16:42:09 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Harold, Said, and I have been trying to put together the motivation/rationale for RuleML. Also working on some applications (personalizable agents). Tue Feb 4 16:42:45 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: We have a big span of issues, rqanging from OWL to databases to .... highly distributed situations we hope to see on SemWeb. Tue Feb 4 16:43:18 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: Last week we decided to have the features document be the RuleML features document. Tue Feb 4 16:43:59 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Could we see some of it next week? Tue Feb 4 16:44:03 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: COuld be done Tue Feb 4 16:44:13 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Prosy meat might take a bit longer Tue Feb 4 16:44:28 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Great -- it'd be good to see a skeleton at least Tue Feb 4 16:44:35 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: SOme more use cases: Tue Feb 4 16:45:07 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Retailing, Business Policies -- executed when someone goes shopping (pricing, recommendations, ad targetting) Tue Feb 4 16:46:40 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Reason from properties of good to configure an auction. (separable goods or not) We did a system with umich, based on rules in contracts, using priority mechanisms to update values in contracts, like reservation price. Add another provision because of some jurisdiction with a special tax, etc. Tue Feb 4 16:48:07 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Authorization or Trust, like TimBL's examples. Are you credit worthy? Fraud Alert Service + Credit Bureau...... Should medical records be shared with a particular doctor, eg because there's evidence of a legitimate cause. Tue Feb 4 16:48:30 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Classic security policy scenarios, eg logic access control. Tue Feb 4 16:49:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: THis is E-Contracting and Business Policy rules. Tue Feb 4 16:49:17 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Let's talk about the structure and size of document. Five or six use cases? Tue Feb 4 16:50:01 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: SO what if some uses are not motivated by any use case? EG configurators have used rules for years. Tue Feb 4 16:50:24 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: It will be hard; I'm just concerned that too many will be confusing/overwhelming. Tue Feb 4 16:51:05 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: I think of Use cases as more in-service to abstract goals. Pick examples that are nice expositorily. We can have annotated pointers to more things. Tue Feb 4 16:51:50 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: What Ben says -- we collect everything first, then we decide which 7+-2 to include. Tue Feb 4 16:52:44 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Harold: (some shopping agent, .......) Tue Feb 4 16:53:04 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: which can understand fine print (delivery charges, etc) Tue Feb 4 16:53:13 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: (I sent the URL) Tue Feb 4 16:54:07 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: Purely declarative rule engine, in Java, with atttachements. Tue Feb 4 16:54:48 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :RACSA http://www.dfki.de/racsa/ Tue Feb 4 16:54:58 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: How do we think about transformation/rewriting rules? Tue Feb 4 16:55:06 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :s/do/should/ Tue Feb 4 16:55:25 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: I think it's critical Tue Feb 4 16:55:55 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: transforming data structures, etc, see my e-mail today. Tue Feb 4 16:56:27 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: You can use rules to map between ontologies without per se "rewriting". Tue Feb 4 16:56:55 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: Rewriting rules dont do rewriting, they do deductive inferernce as any other rules engine Tue Feb 4 16:57:40 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: That's not fair. It's not clear that the best way to do translation is to deductively map everything into a common language. Maybe actually rewriting in place IS better! Tue Feb 4 16:58:17 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: XSLT gives one engineering example -- it's that kind of rewriting tool. It's the default tool for rewriting XML. Tue Feb 4 16:58:28 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: I use it myself! Tue Feb 4 16:58:44 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: But the theory about integrating them is not clear Tue Feb 4 16:59:26 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: THere are systems defined in terms of a directed equation -- a token on the left side which can be rewritten into a body on the right-hand side (a lot like grammar specification). XSLT works similarly. Tue Feb 4 17:00:04 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: A combination of rewriting and throwing-away. Tue Feb 4 17:00:21 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: infer *and* replace in KB Tue Feb 4 17:00:52 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: the deletion is not in the spirit of montonic rules a la horn. There's a fundamental divergence. Tue Feb 4 17:01:11 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Pat: You dont have to think of it as deletion; you can step beside it Tue Feb 4 17:01:25 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Yes, the question is whether we can do that; I prefer it. Tue Feb 4 17:02:20 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Stefan: view it as states. In state 1, we have some things, in state 2 we have some other things, etc. Deduction rules. Deletion is "something is not copied to the next state". Tue Feb 4 17:03:02 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: Yep. Limitted classical negation in LP can help for not moving something to the next state. Integrety constraints are imporetant. Tue Feb 4 17:03:39 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ben: SOme work at ISWC Rules Francois Brie [sp?] Linked from RuleML page Tue Feb 4 17:03:44 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Brye Tue Feb 4 17:04:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :or just Bry Tue Feb 4 17:04:35 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike: Out of time. Next week we hope to have an initial draft of use case document, and maybe get a look at the features doc. Tue Feb 4 17:05:02 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Intersection in OWL Lite. Is it named classes, or named classes and restrictions? Tue Feb 4 17:05:37 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: it's always been in spec, but not overview doc. Tue Feb 4 17:05:43 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: But wasnt it always named classes? Tue Feb 4 17:06:20 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: It's always been there implicitely in the axioms, on the ?-side of the axioms. Tue Feb 4 17:06:39 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: OWL Lite has interesction of named classes and restrictions Tue Feb 4 17:07:08 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: the idfference is that it can only occur in the RIGHT HAND SIDE of the axioms. As is true with restrictions. Tue Feb 4 17:07:39 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Protege and ..., would have trouble with unnamed restreictions. Tue Feb 4 17:08:09 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: I have trouble imagining that. It's like OIL-Ed. Tue Feb 4 17:08:23 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Peter: This view of OWL Like goes way back. Tue Feb 4 17:08:28 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: It's in AS&S Docs Tue Feb 4 17:08:48 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: We had this discussion that it would be left out. Tue Feb 4 17:09:23 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: We agreed to leave it out of the overview, not the language. But now truth takes precidence over presentation. :-) Tue Feb 4 17:11:00 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: I thought OWL Lite only had intersection between named classes. Tue Feb 4 17:12:11 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :...: I had information that Protege did not handle unnamed classes Tue Feb 4 17:12:31 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian: That's a little different from what it has inside. Tue Feb 4 17:12:56 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Yeah, OKBC has it, but just using a GenSym'd class Tue Feb 4 17:13:25 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Peter: But all these things should implement OBKC, and it has unnamed classes in the interface. Tue Feb 4 17:14:13 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Ian; It's been this way in AS&S for a long time. To change it requires raising a real issues, etc. Tue Feb 4 17:14:17 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Deb: Yeah. Tue Feb 4 17:14:30 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike; THanks everyone! Bye! Tue Feb 4 17:14:59 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :========================= Tue Feb 4 17:15:26 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Mike, is logger working? What shall we do with these? Tue Feb 4 17:17:49 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :i'll copy today's portion to the web site and send an email message to joint-committee - thanks very much for scribing today Tue Feb 4 17:18:22 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :My officemate asks if I'll have to keep scribing if I want to keep seeing minutes. :-) Tue Feb 4 17:18:53 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :hmmm - would your officemate be willing to scribe? :-) Tue Feb 4 17:19:48 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :we might trade off between you and me - in the absense of polished minutes, the IRC log is definitely helpful Tue Feb 4 17:20:27 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :My officemate claims he knows nothing of rules or the web or RDF or anything. (and says "Hi, Mike -- ericP") Tue Feb 4 17:20:41 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :although it can also be helpful to get multiple transcriptions Tue Feb 4 17:20:46 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :Hi Eric! Tue Feb 4 17:21:17 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Polished minutes from this kind of discussion would be really hard..... Tue Feb 4 17:21:34 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Much easier to justr write it down without trying to deeply uinderstand the flow. Tue Feb 4 17:22:40 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :agreed - i did that initially (with some success) but it took quite a bit of time and effort Tue Feb 4 17:23:08 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Yeah, and it keeps me awake and paying attention during the meeting. :-] Tue Feb 4 17:23:14 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :anyway, I'm off. 'later! Tue Feb 4 17:23:22 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :quick question - can i embed calculations in cwm rules, e.g. add the values of 2 properties? Tue Feb 4 17:24:28 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :Yes. You can do string concat/regexp type functions too, crypto checking, and lots of cool stuff. Look at swap/test/regression.n3 and follow it to examples. Tue Feb 4 17:24:46 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :I've never actually used that stuff. :-) Tue Feb 4 17:25:11 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :I want to try a dynamic-loading thing, where math:addition is a URI from which you can fetch some python code to do the addition. Tue Feb 4 17:25:19 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu PRIVMSG #joint-committee :anyway, gotta run! Tue Feb 4 17:25:22 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net PRIVMSG #joint-committee :great -- thanks for the pointers and again for scribing today Tue Feb 4 17:25:26 2003: :sandro!-sandro@24-6-251.wireless.lcs.mit.edu QUIT :Client Exiting Tue Feb 4 18:13:59 2003: :mdean!-mdean@w222.z064001205.chi-il.dsl.cnc.net QUIT :