14:54:45 RRSAgent has joined #vmtf 14:54:45 logging to http://www.w3.org/2006/01/24-vmtf-irc 14:54:54 Meeting: SWBPD Vocab Management TF 14:56:12 TBaker has joined #vmtf 14:57:00 SW_BPD(VMTF)10:00AM has now started 14:57:07 +??P16 14:57:22 +Tom_Baker (was ??P16) 14:58:19 q+ to ask Alistair about CVS file structure (2005-11-18/2006-01-18) 14:59:39 Agenda: http://lists.w3.org/Archives/Public/public-swbp-wg/2006Jan/0122.html 14:59:59 RalphS, hi 15:00:04 +Ralph 15:00:46 Previous: 2006-01-17 http://lists.w3.org/Archives/Public/public-swbp-wg/2006Jan/0079.html 15:04:06 zakim, please mute me 15:04:06 sorry, TBaker, I do not see a party named 'TBaker' 15:05:12 zakim, I am Tom_Baker 15:05:12 ok, TBaker, I now associate you with Tom_Baker 15:06:03 zakim, nick TBaker is Tom_Baker 15:06:03 ok, RalphS, I now associate TBaker with Tom_Baker 15:09:22 aliman has joined #vmtf 15:09:30 +??P28 15:09:35 +Alistair_Miles (was ??P28) 15:10:03 zakim, alistair is really me 15:10:09 +aliman; got it 15:13:19 RESOLVED: Ask for WG approval for publication at 6 Feb telecon. 15:13:44 Ralph: main thing standing in way is the IE6 bug 15:14:00 s/bug/not a bug/ 15:14:18 Alistair meant bug in the recipes 15:15:19 Tom: options ... 15:15:30 IE6 does not explicitly request HTML content 15:15:34 (i.e., "clickability" does not work in IE6). Is that a bug 15:15:34 or a feature? Two possibilities: 15:15:34 1) change default content type to HTML (and every RDF toolkit 15:15:34 will have to request RDF content); 15:15:34 2) leave recipes as is and live with lack of clickability 15:15:34 for IE6. 15:15:59 Tom: do we need to have a position and do we need to write anything in the document? 15:16:25 Alistair: if we decide that SemWeb ought to send certain headers then we need to publish that to the community 15:16:34 Alistair: if our message is that "SW apps should send headers", then we need to publicize that 15:16:42 ... and we'd need a registered MIME type for turtle 15:17:08 RalphS: why important to have MIME type for turtle? 15:17:23 s/important/important right now/ 15:17:37 Alistair: if someone decides to publish using turtle but HTML is default content type then SW app 15:17:42 needs some way of requesting the turtle content 15:17:54 RalphS: meant a different qst... 15:18:33 There is no W3C rec specifying turtle, so someone publishing turtle-content only does so at own risk 15:18:46 more a turtle-community issue than this WG... 15:19:24 ...not necessary an issue for this WG... 15:20:53 Sesame's MIME type for turtle: application/x-turtle 15:21:40 RalphS Do we want to make a decision about a default content type in absence of an Accept header? 15:22:13 My position: we would upset too much of SW community if we broke SW apps (i.e., changing the 15:22:18 default to HTML) 15:22:28 RalphS likes option 2 15:23:22 RalphS Not correct to define RDF as the default, but 15:23:37 most technology has behaved as if it was default, and we should not break them. 15:24:32 Solution for now: either say nothing or explicitly acknowledge that there is no default. 15:24:55 Always the case: server does not have to respect the Accept header. 15:25:14 We say explicitly: Server "should" look at the Accept header and return a content type... 15:25:27 In the absence, server MAY choose any content type it wishes. 15:25:51 "As of 2006, we believe there is deployed SW technology that expects..." 15:26:13 Server may use other hints at its disposal - e.g., looking at user agent string. 15:26:54 Tom: one of the early adopters we'd like to win over is Library of Congress for MARC Relator Terms 15:27:13 ... they have an RDF Schema but it's not resolved from the MARC Relator Term URIs at the moment 15:27:54 ... in Ralph's scenario LOC's server could choose to serve HTML as the default behavior as their policy decision 15:29:01 ... i.e. if server SHOULD look at Accept: header and in the absence of Accept: the server chooses the content type 15:29:23 Alistair: the current recipes are all set to serve RDF by default 15:29:48 ... so what can we do to help people who want a different default? 15:30:06 ... apache supports conditional redirects based on user agent field 15:30:19 Alistair: using user agent string, could "work around" the IE6 issue 15:30:47 RalphS: we get flak for recommending evaluation of user agent string 15:33:12 suppose we give advice: "BP is to include Accept header. In absence of headers, here are some tricks..." 15:33:41 if we were to augment the recipes to do redirect on User Agent, how much more complicated would that be? 15:34:10 Alistair: Looking at Recipe 3, it requires an additional two lines 15:34:16 Alistair: consider -> http://www.w3.org/2001/sw/BestPractices/VM/http-examples/2005-11-18/#recipe3 recipe 3 15:34:32 it would be nice to have these examples really "cut-and-paste"! 15:34:57 1+ on really "cut-and-paste" 15:35:21 Alistair: if we want to leave it up to Vocab Owners as to default, could have two lines at top, 2 at bottom, 15:35:28 you comment out the one you want 15:36:21 RalphS If we could have just one place to put "tricks"... 15:38:43 "See Appendix B for a work-around on picking a default based on user agent" 15:40:32 Alistair: how would we get out the message that Accept headers should be sent? 15:40:56 RalphS SWIG - if you don't send an Accept header, at your own risk... 15:41:37 RalphS Someday we might have GRDDL processors getting RDF out of XML, so reasonable to 15:41:47 expect SW apps in future to deal with XHTML, etc... 15:42:04 would not want to paint ourselves too much into a corner by recommending certain content types. 15:42:43 TBaker, you wanted to ask Alistair about CVS file structure (2005-11-18/2006-01-18) 15:42:47 oops 15:44:03 Alistair: previously we were favoring clients ... that don't send accept headers.... 15:44:20 Alistair: more balanced now. Here are the choices, here the consequences. 15:44:43 RalphS We think that flipping this switch can be done in a line or two to comment or uncomment. 15:45:11 ACTION: Add those commented or uncommented lines to recipes in order to flip switch between HTML and RDF by default. 15:45:37 RalphS: Text in cookbook for cut-and-paste purposes -- we should have the RDF default uncommented. 15:45:50 in other words, we should still embed a bias for RDF in the cookbook. 15:46:15 Give them a way to do it other way, but they have to do work for the alternative. 15:46:17 Tom agrees. 15:49:02 Alistair: I'm willing to write "appendix B" showing how to do user agent conditional redirects 15:49:09 TBaker, you wanted to ask Alistair about CVS file structure (2005-11-18/2006-01-18) 15:50:03 This conference is scheduled to end in 10 minutes; all ports must be freed 15:50:13 http://www.w3.org/2001/sw/BestPractices/VM/http-examples/2006-01-18/ 15:51:24 Alistair: ^ new editor's draft work-in-progress 15:52:54 title: Best Practice Recipes for Publishing RDFS and OWL Vocabularies/Ontologies 15:53:05 "Best Practice Recipes for Publishing RDFS and OWL Vocabularies/Ontologies" in http://www.w3.org/2001/sw/BestPractices/VM/http-examples/2006-01-18/ 15:54:28 Alistair: in a complete treatment there will be a recipe 6 with bounded descriptions 15:55:03 This conference is scheduled to end in 5 minutes; all ports must be freed 15:55:12 ... i.e. the server will return just specific information, not information about the whole ontology 15:56:07 ... these vocabularies are distinguished mostly by their size 15:57:37 We recognize that large vocabularies may have special needs so add a placeholder (but not an entire placeholder section). 15:58:01 Alistair: rather not reduce the scope of the document. The missing piece is suggested by the Fifth Recipe. 15:58:01 This conference is scheduled to end in 2 minutes; all ports must be freed 15:58:36 Alistair would be happiest with placeholder - Recipe 6 with a bit of prose. 15:59:07 This conference is scheduled to end in 1 minute; all ports must be freed 16:00:01 This conference is scheduled to end now; all ports must be freed immediately 16:00:03 The time reserved for this conference has been exceeded. 13 ports must be freed 16:03:54 -Ralph 16:03:55 -aliman 16:03:58 -Tom_Baker 16:03:59 SW_BPD(VMTF)10:00AM has ended 16:04:01 Attendees were Tom_Baker, Ralph, Alistair_Miles, aliman 16:04:14 rrsagent, please make this record public 16:25:03 oh, ralph - i added your name to the list of editors for the new editor's draft because I thought you made significant contributions that ought to be acknowledged, let me know however if you'd rather not be named as an editor. 17:52:51 Zakim has left #vmtf 19:22:19 alistair, thanks for including me as an editor but I think you did the bulk of the work and remain as the point person so a simple acknowledgement would be fine 19:22:25 oh, alistair's left 19:22:28 will send him mail 19:22:34 rrsagent, bye 19:22:34 I see 1 open action item saved in http://www.w3.org/2006/01/24-vmtf-actions.rdf : 19:22:34 ACTION: Add those commented or uncommented lines to recipes in order to flip switch between HTML and RDF by default. [1] 19:22:34 recorded in http://www.w3.org/2006/01/24-vmtf-irc#T15-45-11