Session Start: Fri Aug 23 15:19:56 2002 [15:19] *** Now talking in #rdfcore [15:20] zakim, who is here [15:20] gk, you need to end that query with '?' [15:20] zakim, who is here? [15:20] sorry, gk, I don't know what conference this is [15:20] On IRC I see gk, em, Zakim, aaronsw, logger_1 [15:20] zakim, this is SW [15:20] ok, em [15:20] zakim, who is here? [15:20] On the phone I see ??P11, PatrickS, FrankM, ??P14 [15:20] On IRC I see gk, em, Zakim, aaronsw, logger_1 [15:20] *** em changes topic to 'rdfcore 2002-08-23 teleconference' [15:21] +??P15 [15:21] *** DanCon has joined #rdfcore [15:21] *** bwm has joined #rdfcore [15:21] +??P16 [15:22] +??P17 [15:22] zakim, ??P16 is bwm [15:22] +Bwm; got it [15:22] +EricM [15:23] *** em is now known as em-tscrib [15:23] +AaronSw [15:23] *** mags has joined #rdfcore [15:23] zakim, who's here? [15:23] On the phone I see ??P11, PatrickS, FrankM, ??P14, ??P15, Bwm, ??P17, EricM, AaronSw [15:23] On IRC I see mags, bwm, DanCon, gk, em-tscrib, Zakim, aaronsw, logger_1 [15:23] I think ??p14 is GK [15:23] *** mags is now known as danb_lap [15:23] * DanCon Zakim, what's the passcode? [15:23] * Zakim saw 7332 given for the conference code, DanCon [15:23] * danb_lap is running late [15:23] zakim, ??p14 is probably GK [15:23] +GK?; got it [15:23] tscribe == temp scribe... would like to request a additional scribe [15:23] I hear Pat Hayes. [15:24] * em-tscrib will be muted for most of the call [15:24] +DanC [15:24] yes [15:24] yes [15:24] :) [15:24] yes please :) [15:25] err... yes, kind of [15:25] tscribe == temp scribe... would like to request a additional scribe [15:25] ok, lets try it for a while [15:25] zakim, who's here [15:25] bwm, you need to end that query with '?' [15:26] zakim, who's here? [15:26] On the phone I see ??P11, PatrickS, FrankM, GK?, ??P15, Bwm, ??P17, EricM, AaronSw, DanC [15:26] On IRC I see danb_lap, bwm, DanCon, gk, em-tscrib, Zakim, aaronsw, logger_1 [15:26] bwm, can i request an additional scribe and switch for next week? jan are you willing to switcg? [15:26] Pat Hayes [15:26] SteveP [15:26] *** aaronsw is now known as AaronSw [15:26] DaveB, JanG [15:27] bwm, i'd like to request a different scribe please... i'm afraid this wont work [15:27] i'm willing to switch for next week [15:27] uncle! [15:27] efe095uibgf [15:27] fdbn [15:27] gbfgdxsyn [15:27] * em-tscrib damn [15:27] * em-tscrib notes alex says hi [15:27] fdvjr tbbsd [15:27] * DanCon waves to Alex [15:27] :) [15:28] thanks gk [15:28] *** gk is now known as gk-scribe [15:28] *** dajobe-la has joined #rdfcore [15:28] +Hugo_Haas [15:28] zakim, Hugo is DanBri [15:28] +DanBri; got it [15:28] heh [15:29] status of documents ... continued (almost done) [15:29] guha, are you here? [15:30] pat? do you have additional input here? [15:30] http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0168.html [15:31] Minutes, no comments (except GK regrets not noted) [15:31] Completed actions, all OK [15:31] Withdrawn actions, all OK [15:31] Item 9 [15:32] status of documents ... continued (almost done) [15:32] Abstract data model working draft [15:32] working on last pubrules [15:32] Zakim, who's muted? [15:32] I see no one muted [15:32] ipr disclosure and validation [15:32] -??P11 [15:32] almost complete [15:32] eric's muted at his end [15:32] ACtion continues, should be out soon [15:32] (em, right? [15:33] why skip 10? it looked orthogonal to datatypes to me. oh well. [15:33] Item 10 [15:33] yes gk, thanks [15:33] +??P11 [15:33] zakim, ??P11 is PatHayes [15:33] +PatHayes; got it [15:34] DanC, this is item 10 [15:34] * DanCon q+ to suggest deprecating Alt/Seq/Bag [15:34] * Zakim sees DanCon on the speaker queue [15:34] DanC, no it's not -- I didn't see 10 go by -- oh well [15:34] Item 11 [15:35] Problem with rdf:Alt -- different people using it in diufferent ways [15:35] we're not throwing out our existing containers at this stage of the game [15:35] it exists, but we don't know what it does [15:35] especially since daml/owl collection has even less meaning [15:36] DanC, Suggested that dublin core uses alt for disjunction (?) [15:36] people have been asking for a common means of expressing collections for a variety of tasks... i'd rather be more specific/formal declrations for these collections rather than just deprecating these [15:36] the primer should have an appendix for "if you think RDF works like this, please think again" [15:37] DanBri, problem is that people are using alt without creating new vocabulary to go with it [15:38] yes, what danc is saying [15:38] PatH: people are using this such that the value is not the container [15:38] danbri's alt msg http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Aug/0157.html [15:38] DanC: we should write something inthe primer that makes it clear this is wrong [15:38] someone should tell DC fast because it's in the spec they're about to approve [15:38] daveb? [15:38] are you here? [15:38] yes [15:38] Is this an education problem, or more? [15:39] make your own comments on dc rdf docs. I've made mine [15:39] Is the value the container, or something else? What's the point here? [15:39] can you take an action to explain this to dc community as dc/rdf liason? [15:39] i can't explain this vague discussion without a proper descr to point people at [15:39] Current approach is to leave the vocab in existence, but NOT to define the semantics that some people might expect. [15:40] yep, makes sense [15:40] bwm: immediate need is to educate the DC community, prefereably via section in primer [15:41] see http://dublincore.org/documents/2002/04/14/dcq-rdf-xml/ [15:41] Dublin Core document has been sitting and available for 6 months [15:41] ... concern that we have "missed the boat" [15:42] [[ [15:43] 2.2 Repeated Properties and ``MetaMetaData'' [15:43] All Dublin Core Elements are repeatable. Repeating element E: [15:43] The objects mentioned as targets all meet the relation represented by element E with the subject under description. [15:43] In RDF it is also possible to repeat properties with the cited meaning of a logical AND. [15:43] Sometimes one wants to explicitly say something slightly different: The objects fulfil the relation E as a group, in order or picking one or the other doesn't really matter w.r.t. E. [15:43] RDF allows to mark such internal relationships, by creating new resources from given objects in three different ways, called Bag,Seq and Alt construction. [15:43] ]] [15:43] Another typical application of Alt appears with lists of synonyms for a chosen descriptor. As Alt gives a slight precedence to the first given value, this is exactly what happens with synonyms: They form an equivalence class with a chosen representative. [15:43] -- http://dublincore.org/documents/2002/04/14/dcq-rdf-xml/ [15:43] Suggestion that daveb send note to DC explaining that DC use of alt is perceived as "strange"... [15:43] [[ [15:43] Another typical application of Alt appears with lists of synonyms for a chosen descriptor. As Alt gives a slight precedence to the first given value, this is exactly what happens with synonyms: They form an equivalence class with a chosen representative [15:43] ]] [15:43] ... suggest use of rtepeated properties, or... [15:44] ... suggest that they more fully justify use of alt in their own document rather than by reference to RDF spec [15:44] Suggested that rdf:Alt be strongly deprecated, if not removed [15:45] bwm: don't want to reopen this. Can we handle it by dialog? [15:46] What to suggest to them? [15:46] As far as RDF is concerned, the value of the property is the Alt container object [15:46] pls make it clear to them that these Alt semantics would apply only to properties they (dc) define. [15:47] rdf myths! [15:47] em, :-) [15:47] i like that idea [15:48] Model theory says nothing [15:48] ericm, I'd like help [15:48] -action DanC, review primer text and propose alternative [15:49] DanCon, i'm planning on reviewing this anyways [15:49] so consider help, ok [15:49] cool. (you might change your nick) [15:49] *** em-tscrib is now known as em [15:50] -action daveb, talk to DC folks about this issue [15:50] Item 14, datatypes [15:51] PatS: need review of latest document, then go for publication [15:52] DanC: uses rdfs:range; daml restriction not using rdfs:range doesn't give required effect [15:52] PatH: none of these proposals can do "DAML-sensitive datatyping" [15:54] bwm: 3 comments: (1) editors, please be clear about status of document; [15:54] -PatHayes [15:55] ... (2) in reference to untidiness: big noticeable request for community feedback; include test cases to clarify intent (e.g. age "10" and title "10" [15:55] +??P10 [15:55] zakim, ??P10 is Pat Hayes [15:55] ... (3) appendix, please give DC examples in RDF/XML not N-triples [15:55] I don't understand '??P10 is Pat Hayes', AaronSw. Try /msg Zakim help [15:55] zakim, ??P10 is PatHayes [15:55] +PatHayes; got it [15:56] *** danb_lap has quit IRC (EOF From client) [15:56] Noted that literals as subjects is postponed, so examples may need reworking [15:57] FrankM: for tutorial purposes, don't use abbreviations in 1st example (e.g. use of &xsd;) [15:58] ... may introduce these later [15:58] * DanCon q+ to ask if anybody has finished their review of this draft [15:58] * Zakim sees DanCon on the speaker queue [15:59] daveb not done proper review; still waiting for removal of qnames and s/rdf:type/something else/ [15:59] PatS: target for revised document back to WG, next Wednesday (if PatH gets his phone back today) [16:00] PatH: do we need "elementary" introductory material? Some say yes. [16:00] Request for review of doc in current state. [16:02] -action JanG, review datatypes document in current form [16:02] GK will also if time permits [16:02] -action SteveP, review doc in current state [16:02] http://www-nrc.nokia.com/sw/rdf-datatyping.html [16:03] pls send review version to www-archive [16:03] Above is document to review. [16:04] i'm happy with the editors telling us how best they would like reviews [16:04] DanC will review if time permits [16:04] * DanCon wonders if it's clear who's sending the archived copies... the editors? [16:04] kjnbhijxfjknbrkjn [16:04] * em crap [16:04] bwm: no community response to feedback -- none so far [16:04] -PatrickS [16:05] *** danbri has joined #rdfcore [16:05] section 2.3.2? in new doc [16:05] Item 12, rdfms-assertion [16:05] * danbri regains connectivity [16:05] we did this [16:05] http://www.ninebynine.org/wip/RDF-basics/2002-08-05/Overview.htm#section-Social [16:06] danbri, frank, and ? reviewed this i believe [16:06] Proposed to close issue based on text in document [16:06] Who has reviewed this? [16:06] i thought we closed this? [16:07] if it's archived, pls make sure today's record points to it [16:07] -action bwm, check that editors working drafts are properly linked from the WG overview [16:07] * danbri revists http://www.ninebynine.org/wip/RDF-basics/2002-08-05/Overview.htm#section-Social [16:07] [[[Note to scribe, make sure minutes contain link to correct archive message]]] [16:08] no, what i was refering to was somethign we did 4 weeks ago... [16:08] -Bwm [16:08] * DanCon thinks maybe it'll be faster to read it now... [16:08] gk-scribe: please post an uni-zipped copy of the files to wwwarchive, you can't link to a zip's contents [16:08] no bwm is still here [16:08] not on the phone, apparently... [16:09] [[ [16:09] For example, a media type, application/rdf+xml [RDF-MIME-TYPE ] is being registered for indicating the use of RDF/XML that might be published with the intent of being such an assertional representation (as distinguished from other XML or text that may just happen to look like RDF assertions). [16:09] ]] [16:09] Dajobe, I tried but I have software bugs that are preventing that [16:09] err.. is there to be a different doc i should be submitting for publication then? [16:09] when I reviewed this, I asked forthis parag to be ommitted. [16:09] dajobe-la, you can using the new zip: URI scheme, right? [16:09] AaronSw: I think the mime type draft expires soon, need a new one? [16:09] * DanCon q+ to ask about the media type draft [16:09] * Zakim sees DanCon on the speaker queue [16:09] * DanCon q+ to 2nd the motion on the table [16:09] yeah, good point. [16:09] * Zakim sees DanCon on the speaker queue [16:10] ack DanCon [16:10] DanCon, you wanted to suggest deprecating Alt/Seq/Bag and to ask if anybody has finished their review of this draft and to ask about the media type draft and to 2nd the motion on [16:10] ... the table [16:10] * Zakim sees no one on the speaker queue [16:10] my concern: the mime-type text doesn't represent the current content of the mime type doc [16:11] -PatHayes [16:11] FrankM, reviewed in msg of 2 Aug ... concerns with use of triples making distinct assertion ... using the term in two different fashions [16:12] i already pointed tim at this [16:12] -action FrankM, make proposal of alternative text [16:12] hes happy enough for first draft [16:12] no, i made that point clear [16:13] AGREED: issue closed as proposed [16:13] (rdfms-assertion that is) [16:13] graham, do you plan to remove that parag I was concerned about? (re mime type /assertion?) [16:13] the link to the draft is broken [16:14] ack, DanC [16:15] -action aaronsw, reissue MIME type draft [16:15] it expires Aug 30 btw [16:15] 'up where [16:15] '? [16:15] http://www.ietf.org/internet-drafts/draft-swartz-rdfcore-rdfxml-mediatype-00 [16:15] http://www.ninebynine.org/wip/RDF-basics/2002-08-05/Overview.htm#section-Social [16:16] 2.3.2 Social meaning [16:16] of which document? [16:16] Danbri examines currently proposed doc [16:16] the one on ninebynine? [16:16] above url, yes [16:18] the 'social meaning' text works for me cuz it talks about protocols as well as formats. [16:18] * danbri finds the "might be" pretty useless: formally it is weak, but has a stronger subtext [16:19] Note that we will check for consistency between MIME type draft and abstract syntax doc before new draft is published (see action above) [16:19] Oh, DanCon, I like the first paragraph... [16:19] thansk all [16:19] -??P15 [16:19] -??P17 [16:19] -DanC [16:19] Meeting closed [16:19] -EricM [16:19] *** dajobe-la has quit IRC (Client Exiting) [16:19] -AaronSw [16:19] -FrankM [16:19] -GK? [16:19] so the media type spec doesn't need to say 'rdf is for making claims that you can be sued for', but it should say 'if you use this in an HTTP 200 response, you can be held accountable for the claims you make' [16:20] DanC, the problem I see with that is that it puts in the RDF spec... [16:20] It currently doesn't say that. [16:20] i.e. it's not the data format alone, but the data format used in the context of HTTP 200 (and SMTP and ...) [16:20] ... something that is not just about RDF [16:21] I'm not arguing the principle, just how and where to present it. [16:21] er... the WG just agreed to present it there, in 2.3.2, no? [16:21] oh, I misheard, thought there WASN'T after hours talk... [16:21] so the media type spec doesn't need to say 'rdf is for making claims that you can be sued for', but it should say 'if you use this in an HTTP 200 response, you can be held accountable for the claims you make' [16:21] yup [16:22] for some value of 'you' [16:22] * danbri stores false RDF documents on his www server [16:22] hmm... isn't the HTTP spec clear on who the 'you' is? [16:22] zakim, who's here? [16:22] On the phone I see DanBri [16:22] On IRC I see danbri, bwm, DanCon, gk-scribe, em, Zakim, AaronSw, logger_1 [16:22] oh dear, that's not what I heard... what I heard was something like this "RDF used in appropriate supporting context makes statements you can be held accountable for" without actually saying anything about HTTP [16:23] * danbri tries to rejoin [16:23] I expect TimBL to be unsatisifed by anything that doesn't mention HTTP. [16:23] you appear to be the only one on the phone, danbri. [16:23] * danbri tried to reconnect, but it was restricted [16:23] there has to be *some way* to stick RDF in the web in such a way that the specs make it clear that you mean it. [16:23] * danbri thinks its for HTTP to mention HTTP [16:23] Does it need to mention HTTP explicity or only through substitution? [16:24] Dan, this is probably just my ignorance of the HTTP specs, but how does it get from protocol to a notion of 'you', publisher/owner/webmaster etc [16:24] explicitly: dunno. hard to read timbl's mind here. I expect he's argue that it can't hurt to be clear that the relevant contexts include HTTP. [16:24] If one explcictly mentions HTTP, I fear an assumption that other contexts can't imply assertion. [16:24] s/one/RDF/ [16:25] err, resolution of variables (i.e. "appropriate context") [16:25] So I use RDF/XML documents to describe the content of digital images; the situation depicted in some painting, for eg. [16:25] I suppose we could mention HTTP at the level of an example here? [16:25] yes, ", e.g. HTTP" [16:25] or ", e.g. in the context of an HTTP 200 OK response" [16:25] If I publish such an RDF/XML doc by reifying and embedding in another, I'm ok. If I publish it as a separate HTTP resource, I risk being accused of asserting falsehoods.... [16:25] and the rest, about 200 response. [16:26] ...simply cos I run the webserver? [16:26] yes. [16:26] I think the current legal frameworks can establish who is the responsible publisher. [16:26] just like if you had put up english document that says "The president is a clown" [16:26] *** bwm has quit IRC (Ping timeout) [16:26] It may NOT be the ISP running the webserver. [16:27] if you publish slander via HTTP 200, there are tested court cases that that has legal effects, no?> [16:27] heck, there are whole acts on these things [16:27] ... rather than one that says "this potograph shows the president depicted as a clown" [16:27] "if you had put up" is loose [16:27] If a webserver I control published such a doc? [16:27] it could be part of a multi-file entailment test... [16:28] well, there's the safe harbor provision of the DMCA that says if you didn't do any review of the documents you're not liable [16:28] * danbri wonders if PGP-signing is where the line is drawn [16:28] I don't intentionally sign any false RDF documents [16:28] entiailment test: we need to be clear, by way of example.org domains and such, that our tests don't say anything about the real world [16:28] I don't believe anyone's PGP-signed the slanderous or copyright-infringing documents either... [16:28] Hmmm... Joseph Reagle has a paper about attaching statements of intent to XMLDSIGs [16:28] ... (or something like that) [16:29] gk, interesting. I'll have a look at that. [16:29] Let me see if I can find the URI... [16:30] btw... what ever happened to PatH's gobshite example? I liked it, and I saw several endorsements that it should go in our WD. Graham, do you intend to incorproate it? [16:30] I think we can carve things nicely down the middle. (i) RDF/XML docs have propositional content (ii) agents can demonstrate attitdues (believing, fearing, desiring) towards such content. [16:30] the example of "the president is a clown" vs. "this photo shows the president as a clown" example seems worth including too. [16:31] I would be very suprised to be dissaolwed from publishing a wishlist.rdf document on my www site, describing the world as I wish it were. [16:31] 0 etc [16:31] Would write a prose document that says "I own a porche"? [16:32] or, ok, "there are no wars." [16:32] you might, as a form of art. [16:32] Is also bad for papertrail-style apps, as you couldn't keep historical records of past known state (which means they'd need recomputing a whole bunch) [16:32] why not keep historical records? [16:32] it's fine to have old prose documents that say "the president is clinton" [16:32] ... I think this is it: http://www.w3.org/TR/xmldsig-p3p-profile/ [16:33] :danbri :wishes { :world :numberOfWars "0" } . [16:33] ...and publish then via HTTP where they'd be misinterpreted per your rules as the website/publisher claiming they were currently true [16:33] If it's fine to have old prose docs saying that, is it fine to have old RDF/XML docs that do similar? [16:33] yes. [16:33] the whole point is that assertions in RDF are no different from assertions in prose. [16:33] DanCon, re GobShite, I am minded to use it in next rev, but it was noted that the language might need changing for use on a serious W3C public document [16:34] next rev/edit: cool. [16:34] re assertions/prose: yup [16:34] *** gk-scribe is now known as gk [16:34] I have lots of prose on my website that I don't assert, too [16:34] really? [16:35] some of it is unattributed, some is out of date, some is joke or wishlist. [16:35] metadata about that prose (or rdf) is needed before you know who currently puts their name to it. [16:35] you can't jus assume webmaster@that-domain will vouch for all claims made in docs served from that machine [16:35] s/machine/service/ [16:35] no, of course not. [16:36] the "rules", inasmuch as there are any, are on a per-document basis, not a per-site basis. [16:36] the bits I put my name explicitly to, I PGP sign. It shows that I've very likely read it. [16:36] yup, I'm happier with a view in vein of 'this document asserts that President is Clinton' [16:37] look, if I could show in court that you wrote "Dan is a pedophile" and put it on your web site and encouraged lots of people to read it and believe it, I wouldn't need a PGP signature to sue you. [16:37] aaron, re: :danbri :wishes { :world :numberOfWars "0" } . I could argue that has implicit reiofication, or something that is not pure RDF. [16:37] not pure RDF: quite! huge requirement not met! [16:37] timbl has always been annoyed that RDF 1.0 doesn't have real quoting. [16:38] -DanBri [16:38] SW_RDFCore()10:00AM has ended [16:38] * DanCon has an appointment at 11amCT, must dash [16:39] Well, our charter didn't really provide for us to add that. We've touched on this from time to time but no real argument to include it was made. I've tended to assume it's a high-prio item for V2. [16:40] Oh well, off to write up the minutes (yawn). Session Close: Fri Aug 23 16:44:14 2002