15:50:50 RRSAgent has joined #webont 15:55:59 sandro has joined #webont 15:56:28 DanC has changed the topic to: 29 May http://www.w3.org/2001/sw/WebOnt/ chair: JimH; scribe: Sandro 15:57:34 baget has joined #webont 15:57:54 agenda + 29May http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0366.html 15:59:32 SW_WebOnt()12:00PM has now started 15:59:36 +Tim_Finin 16:00:13 zakim, what is the code? 16:00:13 the conference code is 9326, sandro 16:00:24 +Evan_Wallace 16:00:42 +Sandro 16:01:34 +[INRIA] 16:02:07 +??P22 16:02:11 +DanC 16:02:15 jjc has joined #webont 16:02:25 +??P24 16:02:43 +??P25 16:03:02 +??P26 16:03:18 +??P27 16:03:53 Zakim, Tim_Finin is JeffH 16:03:53 +JeffH; got it 16:04:08 +[UMD] 16:04:08 +??P29 16:04:09 Jhendler has joined #webont 16:04:11 pfps has joined #webont 16:04:16 Zakim, ?P29 is jjc 16:04:16 sorry, jjc, I do not recognize a party named '?P29' 16:04:25 Zakim, ??P29 is jjc 16:04:25 +jjc; got it 16:04:33 zakim, [umd] is hendler 16:04:33 +hendler; got it 16:05:04 Zakim, ??P22 is MasahiroH 16:05:04 +MasahiroH; got it 16:05:25 ack sandro 16:05:44 RRSAgent, pointer? 16:05:44 See http://www.w3.org/2003/05/29-webont-irc#T16-05-44 16:06:14 Jim convenes Meeting 16:06:43 Evan_Wallace regrets for last meeting missed. 16:06:44 nmg has joined #webont 16:07:19 (continued from 8 May, in particular) 16:07:20 aciton from last week, on DanC, imports test, should have been "Continued" 16:07:37 RESOLVED: ammended minutes from last week accepted. 16:07:45 Action Review 16:07:56 +??P30 16:08:05 I did some work on integrating implementation feedback/news into http://www.w3.org/2001/sw/WebOnt/rqim.html 16:08:15 ACTION: Jim Hendler to report back on implementation feedback. 16:08:25 cont 16:08:26 ACTION: Jim Hendler and Dan Connolly to compose a response to the LC 16:08:26 comment on local and compound keys. 16:08:27 cont 16:08:34 ACTION: Peter Patel-Schneider to work on proof for issue 5.26 B1,B2 16:08:41 moved to JJC 16:08:50 ACTION: Dan Connolly publish XML syntax 16:09:04 discussion: 16:09:17 DanC: I asked for a change to the abstract. 16:09:37 MH: that's fine 16:10:11 MH: 16:10:17 DanC: either way is fine 16:10:25 ACTION: Peter Patel-Schneider and Jeremy Carroll to discuss solutions to 16:10:32 DONE, to be discussed today 16:10:37 ACTION: Jeremy Carroll to make changes to test document concerning 16:10:38 DONE 16:10:52 ACTION: Sandro Hawke to do publishing of Test 16:10:58 CONTINUED 16:11:13 title page date is 28 may, close date is 27 june 16:11:32 DanC: add i18n to your peers list. 16:12:32 C.4.2. Concerning rdf:XMLLiteral http://www.w3.org/2002/03owlt/editors-draft/draft/proposed-misc-200-xmlliteral#proposed-misc-200-xmlliteral 16:12:35 ACTION: Mike Smith and Guus Schreiber will review proposed changes to 16:12:44 ^perhaps that obviates the need for mail, jim? 16:12:50 JimH: I dont understand this option. Mike and Guus not here. 16:12:54 CONTINUED 16:13:05 ACTION: Jeff will post proposed response 16:13:21 DONE 16:13:23 ACTION: Mike Smith will pdate issue list 5.06 (in re: imports closure) 16:13:24 DONE 16:13:31 ACTION: Mike Smith to update issue list 5.20 about synonyms 16:13:37 CONTINUED 16:14:20 ACTION: Jos de Roo to propose a test based on the example given by peter 16:14:22 CONTINUED 16:14:28 +PatH 16:14:31 ACTION: Guus to propose a definition for "class description". 16:14:33 CONTINUED 16:14:49 ACTION: Mike Smith to update the Guide glossary accordingly 16:15:31 ACTION DanC: get I18N WG review of last call specs, continued from ~17 years back 1/2;-) 16:16:34 JimH: I thought there was a response from Martin.... I can't find it right now. 16:17:02 DanC: lets examine all LC comments today 16:17:19 JimH: Next telecon is 5 June, normal time and place 16:17:23 Zakim, pick a scribe 16:17:23 Not knowing who is chairing or who scribed recently, I propose ??P30 16:17:29 Zakim, pick a scribe 16:17:29 Not knowing who is chairing or who scribed recently, I propose jjc 16:17:32 Zakim, pick a scribe 16:17:32 Not knowing who is chairing or who scribed recently, I propose Evan_Wallace 16:17:46 Zakim, who is on the phone? 16:17:46 On the phone I see JeffH, Evan_Wallace, Sandro, [INRIA], MasahiroH, DanC, NickG, ZivH, IanH, PeterFPS, hendler, jjc, ??P30, PatH 16:18:09 Zakim, ??P30 is CharlesW 16:18:09 +CharlesW; got it 16:18:20 Zakim, pick a scribe 16:18:20 Not knowing who is chairing or who scribed recently, I propose [INRIA] 16:18:24 Zakim, pick a scribe 16:18:24 Not knowing who is chairing or who scribed recently, I propose PeterFPS 16:18:28 Charles volunteers to scribe in TWO WEEKS. 16:18:36 Next Scribe: PFPS 16:18:58 JimH: Half of our last call issues overlap the "beer convo" 16:19:04 Zakim, who's talking? 16:19:04 zakim, agenda? 16:19:04 I see 1 item remaining on the agenda: 16:19:05 1. 29May http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0366.html [from DanC] 16:19:13 http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0349.html 16:19:15 DanC, listening for 10 seconds I heard sound from the following: Sandro (5%), MasahiroH (29%), hendler (39%), CharlesW (4%) 16:19:31 AGENDUM Review of "Beer conversation decisions" 16:19:44 http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0349.html 16:20:03 JJC: we talked about 8 issues, of which the last was casual 16:20:13 ... on all of them we were seeing eye-to-eye 16:20:38 ... A) intersectionOf semantics 16:20:48 ... msg 116 16:20:54 ... started as rdf-logic thread 16:21:10 ... mismatch between OWL Full semantics and OWL DL semantics 16:21:29 ... Peter described two possible fixes; we agreed one was better 16:22:08 ... fix #2 is only an if --> iff & correspondence proof. fix #1 was broader 16:22:12 ... sample test case: 16:22:27 ... (prefered one is #2) 16:22:49 ... one of John, Mary, one of John, one of Mary; bigger class is union of two smaller classes 16:23:01 ... currently that doesn't follow under RDFS semantics 16:23:11 PH: really?! 16:23:20 PH: Oh, under OWL Full, okay. 16:23:51 PFPS: right now Intersection, etc, are syntax. They only have necessary, not necessary and sufficient conditions. 16:24:05 PH, JJC: we don't find the "syntax" part of that comment useful 16:24:21 JJC: so one line change to AS&S 16:24:32 DanC: section 5.2 16:24:39 PH: that sounds like a bug 16:25:02 PFPS: it's coherent this way. this was the original intent 16:25:23 JimH: does anyone have a problem with this decision? 16:25:50 PFPS: (laughing) this way I have to do all the work! 16:26:31 JimH: straw pole on accepting solution #2 from message 116 ? 16:26:40 JimH calls the question 16:27:09 JJC: not just IntersectionOf, but all the relevant connectives 16:27:14 msg 116 http://lists.w3.org/Archives/Public/www-webont-wg/2003May/0116.html 16:27:22 JimH hears no objections or abstentions. 16:28:07 JJC: relevent ones are clearly linked in semantics 16:28:10 RESOLVED 16:28:21 DanC: is there a test case? Please add it. 16:28:55 ACTION: jjc {a,b} unionOf {a} {b} 16:29:04 ACTION Peter Change the documents to reflect this #2 in 116 change 16:29:09 ACTION: Peter Change the documents to reflect this #2 in 116 change 16:29:31 JJC: B). B1, B2 in OWL DL Syntax 16:29:42 seanb has joined #webont 16:29:59 ... I've got a sketh proof on a scrap of paper; this is a lot easier after the above fix is done 16:30:41 ACTION: jjc Send proof sketch of extension to correspondence theorem with B1 16:31:08 ACTION: pfps review proof sketch of correspondence with B1 B2 16:31:36 JimH: I think the WG already agreed we wanted this done. All set. 16:31:43 JJC: C) rdf:LIst decision by RDF Core 16:32:12 ... I agreed with Peter that this could be addressed by making the list triple optional in mapping rule 16:33:32 [ missed a bit ] 16:33:54 ... We think there's an answer here, but I can't promise it 16:34:06 JimH: Would Sean's work be affected? 16:34:11 JJC: slightly 16:34:56 ... "to recognize an unnamed class, look for such-and-such a triple", with the rdf:type triple always optional 16:35:10 DanC: I'd like to hear from someone building a syntax checker. 16:35:27 JimH: Not clear to me how this gets exposed in RDF/XML syntax 16:35:57 zakim, who is talking? 16:36:00 JJC: It means that you could use rdf:Description in places you might otherwise have owl:Foo 16:36:08 Jhendler, listening for 10 seconds I heard sound from the following: Sandro (20%), MasahiroH (4%), PeterFPS (15%), hendler (35%), jjc (44%), CharlesW (4%) 16:36:39 JJC: it would be interesting to hear from Sean 16:36:44 Ian: I can ask him tomorrow 16:37:36 JimH: No opposition to JCC's proposed flowchart ("proof works? then... ") 16:37:58 Ian steps out to grab Sean 16:38:26 JimH: feed back to RDF Core? 16:38:46 DanC: not now at least 16:39:07 JJC: D) OWL DL syntax NP complete #83 and #52 from may 16:39:13 ... april #3 16:39:31 ... mapping rule for owl equiv classes embeds hamiltonian path problem 16:39:50 ... equivClass should instead map to 16:40:06 ... "arbitrary connected graphs" 16:40:29 PH scares everyone with joke 16:41:25 DanC: this falls below my threshhold 16:41:28 ACTION: jjc equivalentClass with 4 components and no hamiltonian path 16:41:52 test case will be one which is connected but with no Ham. path. 16:42:26 +??P0 16:42:42 Zakim, ??P0 is SeanB 16:42:42 +SeanB; got it 16:43:08 JJC: To change the mapping rule as in #349(D) 16:43:12 2nd 16:43:24 Any objections? 16:43:26 Any Abstain 16:43:29 RESOLVED 16:43:57 ACTION: Peter put #349(D) change into S&AS 16:44:18 Back to rdf:List issue, with Sean 16:44:51 JJC: can we get rid of type node on all blank nodes? I think so, but haven't proved it yet. 16:45:12 ... Does this make your syntax checker harder? 16:45:37 Sean: So you recognize using first/rest, instead of type rdf:List ? I don't think that would be a problem 16:45:53 JJC: how about owl:Restriction, ... etc 16:45:58 Sean: that's probably okay 16:46:19 Sean: the type arc makes it a little easier, but it should still be easy to do 16:46:39 Ian: How far is this allowed to stretch? To all RDF entailment? 16:46:57 ... eg subproperties of onproperty 16:47:03 JJC: No, that's outside of DL 16:47:17 PH: one application of an rdfs:range 16:47:31 DanC: this should not be looked at as entailment -- just graph stuff 16:47:51 JimH: We might need a table of what you check for using what other things. 16:47:56 JJC: that would be in the proof 16:48:12 JimH: But we'd like something more easy to read, for implementors 16:48:49 DanC: We had a proposal for putting in our document instructions on how to put the toothpaste back in the tube, but we decided not to do that. 16:49:05 JimH: As an implementor, I want this to be easy to find. Rules, whatever 16:49:15 DanC: I can't separate these issues 16:49:37 JJC: The triples tables, from the formal grammar, automatically, .... 16:49:57 JimH: Not expecting people to mine doc for the corresponoding unique feature 16:50:12 DanC: We are in the unhappy state right now. 16:50:41 DanC: We don't have a section which says: "Suppose you have an RDF graph and you want to make an AS out of it, here's what you do: ..." 16:50:56 DanC: we only specify the rules AS->RDF 16:51:23 JJC: Sean's test on this is really quite good 16:51:40 ... We have two OWL syntax checkers already, so maybe that's good enough 16:51:49 DanC: Yeah, that's why this was/is okay with me. 16:51:56 (hmm... add sean's text to the reference doc? i.e. non-normative, but pretty useful) 16:52:45 JJC: WG decision waits for JJC's proof and Peter and Sean reviewing it. 16:53:11 JimH: anyone likely to object then? (hears no comment) 16:53:44 http://www.w3.org/2002/12/open-issues?ml=public-webont-comments&range=1&realm=Public 16:54:40 DanC: I advise going through LC list now 16:54:52 JimH: let's get through JJC's stuff first, quickly. 16:55:09 JJC: E) unnamed individuals 16:55:22 ... Cycles in blank nodes for unname indivs, etc. 16:55:38 ... Rationale is plausible, but it's too much work at this point 16:55:47 ... So we want to "postpone" 16:55:59 JimH,DanC: but we need to create the issue to Postpone 16:57:05 E) unnamed individuals 16:57:05 http://lists.w3.org/Archives/Public/public-webont-comments/2003May/0052.html 16:57:05 owlsas-rdfcore-bnodes-restrictions 16:57:13 Peter and I propose we postpone this issue. 16:57:13 Rationale: 16:57:13 - the amount of work to do otherwise is large, and is not guaranteed to 16:57:13 succeed. 16:57:13 - however the point is a good one and should be reviewed once the community 16:57:15 has more experience. 16:57:21 -- JJC #349 16:57:46 PH: not treated as existentials?? 16:58:11 (it's now clear to me that we can't do this quickly; somebody should raise the issue and somebody should propose to postpone it) 16:58:19 PFPS: I have a niggling worry that it will be more difficult if we allow cycles. To get rid of that worry, it'll be a lot of work. 16:58:26 ... I need a proof. 16:58:35 JJC: I concur 16:59:22 JimH: I hear the group generally willing to postpone on this 16:59:58 JimH: Pat can read when you (JJC) submits to to the WG 17:00:30 ACTION JJC: propose response to owlsas-rdfcore-bnodes-restrictions, including new issue and proposal to postpone 17:00:52 JJC: F) ambiguity 17:01:19 ... comments from RDF Core members, but not officially RDF Core WG 17:01:53 DanC: the term "Non-deterministic" should not be scary here 17:02:09 JJC: just editorial 17:02:32 DanC: Graham's said his was not a real issue 17:02:56 DanC: editor clarify as he sees fit 17:03:22 (has a proposed reply been written....? no one knows. :-) 17:03:30 Apologies for dipping in and out, but I have to leave... 17:03:47 JimH: Peter, if you wish to make editorial changes in clearing this up, please go ahead. 17:04:03 JimH: Sean's reverse-mapping document... 17:04:14 ... we'll review, and we'll decide whether to add it 17:04:22 Sean: should have more in a couple days 17:04:24 -SeanB 17:05:10 JJC: G) OntologyProperty 17:05:16 ... this looks like a bug 17:05:51 ... How is it different from an AnnotationProperty? OntologyProperty has built in domain/range constraints. 17:06:24 JJC: I;m okay with just deleteing the rdf type OntologyProperty 17:06:31 Peter: I don't care 17:07:04 JimH: 2 proposals, (1) drop OP, (2) fix it 17:07:25 JJC: I thought there was an issue, but I couldnt find it. 17:07:52 JimH: does this bother you Jeff? 17:08:17 Jeff: I think Peter came up with OP to express something we needed at one point. I don't particularly need it. 17:09:16 JimH: let's do the best thing for the language 17:09:54 JJC: the concept of OP is kind of in S&AS already, so that's why it's named.... 17:10:10 JimH: I can see OP being useful 17:11:13 JimH: PROPOSED: do 349(G) "peter and I propose that we add an ontologyproperty directive which mirrors the annotationproperty directive in the abstract syntax, with appropriate mapping rules." 17:11:36 JimH: any objections? 17:11:44 JimH: abstentions? 17:11:47 PH, DanC 17:12:12 RESOLVED 17:12:54 JimH: Sean has proposed backwards mapping.... 17:13:02 DanC: which doc would it end up in? 17:13:28 JimH: Reference, as appendix, would be my guess. 17:14:11 ACTION: DanC and Jeremy to review Sean's text for non-normative inclusion into some OWL document 17:14:49 =============== 17:15:57 JJC: pub of TEST is now DONE 17:16:39 JimH: let's postpone the agenda item about presenting owl, it's being dicussed on list. 17:16:41 ============== 17:16:59 DanC goes through LC comments 17:17:18 ACTION DanC: bring up QA review of OWL Semantics with WG 17:18:09 Zakim, please pick a victim 17:18:09 Not knowing who is chairing or who scribed recently, I propose Evan_Wallace 17:18:11 Zakim, please pick a victim 17:18:11 Not knowing who is chairing or who scribed recently, I propose ZivH 17:18:25 Zakim, please pick a victim 17:18:25 Not knowing who is chairing or who scribed recently, I propose PeterFPS 17:18:27 Zakim, please pick a victim 17:18:27 Not knowing who is chairing or who scribed recently, I propose Evan_Wallace 17:18:46 Zakim, please pick a victim 17:18:46 Not knowing who is chairing or who scribed recently, I propose DanC 17:18:48 Zakim, please pick a victim 17:18:48 Not knowing who is chairing or who scribed recently, I propose [INRIA] 17:19:01 Date: Thu, 1 May 2003 22:12:21 -0400 (EDT) 17:19:01 From: "Jimmy Cerra" 17:19:01 To: public-webont-comments@w3.org 17:19:01 Message-ID: <000001c31050$3becd230$0100a8c0@drsoong> 17:19:01 Subject: rdf:about and owl:sameIndividualAs oddities 17:19:21 ACTION PatH: to follow up with that one, it's an educational issue 17:19:52 Anton one closed, though it doesnt show that way 17:20:00 Lacy is happy 17:20:07 Privacy in OWL... 17:20:27 [ most people, including scribe cannot see issues list DanC is reading; w3 machines having problems. ] 17:20:35 JimH: that's on my list 17:20:45 DanC: published subjects 17:20:50 JimH: I'll do that one 17:22:32 DanC: Is AS&S in error about what's DL and what isnt? 17:22:42 Peter: WG decisions are silent about what's in DL 17:22:54 DanC: sameAs should be in DL 17:23:00 JimH: Second 17:23:41 JimH: Peter, please reply to Guus if you're not going to do it. 17:24:23 DanC: let's try to close in May as many as we can, to get more help from machine. 17:25:03 ACTION JimH: prepare definitive list of LC issues for next telecon 17:25:36 -Evan_Wallace 17:25:37 -PeterFPS 17:25:38 -PatH 17:25:40 ADJOURNED 17:25:41 -IanH 17:25:42 -[INRIA] 17:25:43 -JeffH 17:25:44 -jjc 17:25:45 -ZivH 17:25:45 -MasahiroH 17:25:47 -NickG 17:26:54 Jim: I like actions separated 17:27:05 Dan: minutes as sentences 17:27:22 Jim: pointer to IRC for actions, decisions. 17:27:36 IRC okay for blow-by-blow. 17:28:31 decisions clean 17:29:08 plaintext required for minutes, HTML supplement okay 17:29:13 -hendler 17:29:17 -CharlesW 17:29:23 -DanC 17:29:24 -Sandro 17:29:25 SW_WebOnt()12:00PM has ended 17:29:45 RRSAgent, help 17:30:29 RRSAgent, bye 17:30:29 I see 26 open action items: 17:30:29 ACTION: Jim Hendler to report back on implementation feedback. [1] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-08-15 17:30:29 ACTION: Jim Hendler and Dan Connolly to compose a response to the LC [2] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-08-26 17:30:29 ACTION: Peter Patel-Schneider to work on proof for issue 5.26 B1,B2 [3] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-08-34 17:30:29 ACTION: Dan Connolly publish XML syntax [4] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-08-50 17:30:29 ACTION: Peter Patel-Schneider and Jeremy Carroll to discuss solutions to [5] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-10-25 17:30:29 ACTION: Jeremy Carroll to make changes to test document concerning [6] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-10-37 17:30:29 ACTION: Sandro Hawke to do publishing of Test [7] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-10-52 17:30:29 ACTION: Mike Smith and Guus Schreiber will review proposed changes to [8] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-12-35 17:30:29 ACTION: Jeff will post proposed response [9] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-13-05 17:30:29 ACTION: Mike Smith will pdate issue list 5.06 (in re: imports closure) [10] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-13-23 17:30:29 ACTION: Mike Smith to update issue list 5.20 about synonyms [11] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-13-31 17:30:29 ACTION: Jos de Roo to propose a test based on the example given by peter [12] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-14-20 17:30:29 ACTION: Guus to propose a definition for "class description". [13] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-14-31 17:30:29 ACTION: Mike Smith to update the Guide glossary accordingly [14] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-14-49 17:30:29 ACTION: DanC to get I18N WG review of last call specs, continued from ~17 years back 1/2;-) [15] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-15-31 17:30:29 ACTION: jjc {a,b} unionOf {a} {b} [16] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-28-55 17:30:29 ACTION: Peter Change the documents to reflect this #2 in 116 change [17] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-29-09 17:30:29 ACTION: jjc Send proof sketch of extension to correspondence theorem with B1 [18] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-30-41 17:30:29 ACTION: pfps review proof sketch of correspondence with B1 B2 [19] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-31-08 17:30:29 ACTION: jjc equivalentClass with 4 components and no hamiltonian path [20] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-41-28 17:30:29 ACTION: Peter put #349(D) change into S&AS [21] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T16-43-57 17:30:29 ACTION: JJC to propose response to owlsas-rdfcore-bnodes-restrictions, including new issue and proposal to postpone [22] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T17-00-30 17:30:29 ACTION: DanC and Jeremy to review Sean's text for non-normative inclusion into some OWL document [23] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T17-14-11 17:30:29 ACTION: DanC to bring up QA review of OWL Semantics with WG [24] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T17-17-18 17:30:29 ACTION: PatH to to follow up with that one, it's an educational issue [25] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T17-19-21 17:30:29 ACTION: JimH to prepare definitive list of LC issues for next telecon [26] 17:30:29 recorded in http://www.w3.org/2003/05/29-webont-irc#T17-25-03