17:06:28 RRSAgent has joined #tagmem 17:06:28 logging to http://www.w3.org/2011/10/20-tagmem-irc 17:06:45 Not convening, no quorum 17:07:33 will still talk about things anyway 17:10:14 W3C likes to use URIs for extensibility... dissonance is that it's hard to get a permanent URI 17:10:34 sometimes the URI sytnax doesn't fit, or the registry names are htought to be shorter 17:10:56 sometimes you want documentation, a need for central control, some kind of review etc. 17:11:26 not sure 'central control' is the real issue. it's more about trust. decentralized means untrustworthy 17:11:43 Does the hapiana discussion know about the example of the XPointer scheme registry 17:11:47 ? 17:12:21 To what extent is an XML namespace a registry 17:13:14 The XHTML namespace document is used sort of like a registry 17:14:39 Registery + URI as fallback as an issue 17:15:18 could content-type also be a URI? 17:15:49 There may be a lesson from the history of the link-relation registry... started out as URI-backed, ended up not 17:16:22 Why didn't link-rel work, remind us? 17:17:31 http://www.w3.org/2005/04/xpointer-schemes/ 17:17:57 (d) suggests an advantage to URI-backed 17:19:04 case sensitivity might also be an issue 17:19:15 persistence & permanence might have also been an issue 17:19:18 The crucial point about the XPointer registry is that there is a functional mapping from (registered) scheme name to a URI, which can be dereferenced to get to the registry entry 17:19:46 http://www.w3.org/2001/tag/2002/0129-mime is obsolete 17:19:59 And it does seem to have worked, still be working, in contrast to the similar link-rel example 17:21:53 is there a way of getting old findings marked about their status? 17:22:00 superseded by webarch 17:22:16 ht: should be marked as superseded 17:23:13 Henry will raise the issue of editing the document 17:23:23 Superseded by webarch and Authoritative Metadata 17:23:35 http://www.w3.org/2002/06/registering-mediatype 17:24:14 revised 6/21/2011 by Philippe 17:26:11 Including a reference to how an XML namespace is like a registry 17:26:46 http://tools.ietf.org/html/draft-iab-extension-recs 17:26:46 17:27:17 The W3C media type document refers to the April 2004 TAG finding... that may be how LM found it (HT said it's not in the list of TAG findings & wondered how LM found it) 17:28:59 The string 'URI' does not occur in the IAB draft 17:29:54 very little on registries in recs-07 17:32:43 also https://plus.google.com/#103429767916333774260/posts/R6dPzhbc94R 17:33:21 +TimBL 17:33:50 rrsagent, pointer? 17:33:50 See http://www.w3.org/2011/10/20-tagmem-irc#T17-33-50 17:33:56 rrsagent, make logs public 17:37:41 https://plus.google.com/#103429767916333774260/posts/R6dPzhbc94R 17:38:26 http://infrequently.org/2011/09/things-the-w3c-should-stop-doing/ 17:38:39 maybe https://plus.google.com/103429767916333774260/posts/R6dPzhbc94R ? 17:38:45 (no #) 17:39:06 Larry Masinter - "The tag-soup+recovery model of HTML5 is barely scalable" The way in which 'tag soup' doesn't scale is not in the number of instantiations of implementations of receivers (sure, there are lots of instances of multiple browsers), but rather in the number and diversity of implementations. If there are only a handful of browser implementations, you might get some kind of agreement among the 'major browsers', but we should 17:39:06 be designing the language to support hundreds, thousands of implementations. And the more complex the rules are for receivers to be compatible with tag soup, the more likely it is that scalability in that dimension will be hampered. 17:40:29 "it's actually easier to implement an HTML parser [than an XML parser]" - IH 17:41:39 Scheme was optimized for consumer diversity (probably 30-50 implementations), Common Lisp for producer diversity (many more programmers) 17:43:43 lm: Precision is not an unmitigated advantage 17:47:04 HT's email today: http://lists.w3.org/Archives/Public/www-tag/2011Oct/0068.html 17:47:07 -------------------------------------------- 17:51:46 How then addresses the clash of anchor and rdfa id? 17:56:04 ls 17:57:04 in general i'm not happy about using fragment identifiers for anything other than identifying fragments of documents 17:57:18 and I recognize that RDF doesn't do that 17:57:38 and I think that using markup would be better, etc. 17:58:53 so having an opinion on this requires me swallowing things that i don't really believe in 17:59:29 wow 18:00:00 -ht 18:01:44 simple markup like 18:01:49 s p o 18:01:51 instead of xxxx#yyyy 18:02:59 Tim's theory: for each URI xxxx, there is a namespace NS(xxxx). Then xxxx#yyyy means yyyy, according to NS(xxxx). 18:03:33 ? 18:04:05 or even a new URI scheme concept:xxxxencoded#yyyyy 18:04:27 like TDB 18:04:40 maybe tdb allows fragment identifiers? 18:04:59 you're making hypetext have special stauus among apps on the web 18:05:01 because tdb never really returns a document anyway 18:06:44 http://example.com/libs/time#Date 18:07:02 ---> js Date = function(,... 18:09:05 BRB 18:19:33 TTML = http://www.w3.org/TR/ttaf1-dfxp/ 18:22:32 " The scope should be different than that of any other Community Group (but it may be the same, such as when two communities wish to explore two solutions to the same set of problems)." 18:22:57 http://www.w3.org/community/about/agreements/ 18:23:21 -Masinter 18:23:23 -TimBL 18:23:24 -Jonathan_Rees 18:23:25 TAG_Weekly()1:00PM has ended 18:23:27 Attendees were Masinter, ht, Jonathan_Rees, TimBL 18:23:45 rrsagent, pointer 18:23:45 See http://www.w3.org/2011/10/20-tagmem-irc#T18-23-45 19:02:37 Zakim has left #tagmem 20:01:31 timbl has joined #tagmem