01:27:52 noah has joined #tagmem 02:08:43 timbl has joined #tagmem 03:16:58 DanC_ has joined #tagmem 14:15:45 RRSAgent has joined #tagmem 14:15:45 logging to http://www.w3.org/2009/12/09-tagmem-irc 14:15:55 ScribeNick: ht 14:15:59 Scribe: Henry S. Thompson 14:16:07 Chair: Noah Mendelsohn 14:16:16 Meeting: TAG f2f 14:16:37 Agenda: http://www.w3.org/2001/tag/2009/12/08-agenda.html 14:17:47 Present: Tim Berners-Lee, Dan Connolly, Ashok Malhotra, Larry Masinter, Noah Mendelsohn, Jonathan Rees, Henry S. Thompson 14:17:57 [Agenda planning. . .] 14:18:38 NM: Let's try issue HttpRedirections-57 14:18:41 agenda order is 6, 17 14:19:16 Present+ John Kemp 14:19:21 agenda order is 6, 10, 17 14:19:45 Topic: Metadata Architecture (HTTP Semantics): ISSUE-57 (HttpRedirections-57): The use of HTTP Redirection 14:19:51 http://www.w3.org/2001/tag/2009/12/08-agenda.html#HttpRedire 14:20:34 Zakim, remind us in 40 minutes to wrap up redirections and move to versioned specs 14:20:34 ok, DanC_ 14:20:35 Ashok has joined #tagmem 14:21:17 JR: http://lists.w3.org/Archives/Public/www-tag/2009Jun/0057.html 14:21:49 JR: Going through the history---first two points are the origin of this 14:21:56 masinter has joined #tagmem 14:22:13 JR: 1) 303s aren't supposed to be cached -- bug in 2616 -- fixed in HTTPbis 14:22:24 DC: Let's endorse that fix 14:22:53 LM: Not sure about that -- not prepared to endorse -- abstain 14:23:42 NM: This becomes relevant because we encouraged people to _use_ 303 14:24:08 JR: Any reason not cache 303 responses? 14:24:10 LM: No 14:24:50 NM: draft RESOLUTION: TAG endorses the proposed change to HTTPbis to allow caching of 303 responses 14:25:30 DC: Specific proposal is where? 14:25:44 http://tools.ietf.org/html/draft-ietf-httpbis-p2-semantics-08#section-8.3.4 14:25:44 johnk has joined #tagmem 14:26:06 is this OK? "A 303 response SHOULD NOT be cached unless it is indicated as 14:26:06 cacheable by Cache-Control or Expires header fields." 14:26:42 JR: This is different from 307. . . 14:27:09 DC: I think the HTTP spec. is usually neutral wrt caching 14:27:29 JR: OK, we need to explore this further -- the difference from 307 is worrying 14:27:38 I heard DC say HTTP was neutral in the absence of cache-control or expires header 14:27:40 ACTION jonathan: research 303 caching change in HTTPbis 14:27:40 Created ACTION-347 - Research 303 caching change in HTTPbis [on Jonathan Rees - due 2009-12-16]. 14:28:50 JR: Sub-issue 2) There's a need for a non-3xx response, in order that the original URI stays in the status bar 14:29:23 ... Unlike 302 or 307, where the target goes in the address bar 14:29:35 (researching the bug...) 14:29:37 ... This is described as a security concern 14:31:31 s/302/302, 303/ 14:31:43 (many/most purl users want the purl bookmarked, not the redirected addressed) 14:32:59 TBL: But we really don't want that for e.g. 307, because it's only a _temporary_ redirect, so people shouldn't e.g. bookmark it 14:33:16 LM: The single result is insufficient for what we want to tell the user 14:33:29 s/result/result display in the address bar/ 14:33:44 LM: Doing UI design is inappropriate for us. . . 14:34:00 JR: I agree, that's why I want to lose this part of the issue 14:34:31 LM: The principle we can endorse is that the URI you see should be a URI you can use to get you what you see 14:35:00 ... Going further to say it should be a long-term, bookmarkable, etc. URI is a bit fuzzier 14:35:26 NM: WebArch says use one URI for a resource 14:36:26 ... even when they're not going away, it can be a problem, for example when example.com redirects to example-1.com or example-2.com for load balancing 14:36:40 JR: What should I do 14:37:02 For all practical purposes it's impossible to get a purl.org URI into your bookmarks list 14:37:19 DC: Let's find out why Mozilla decline to address the PURL folks' request to fix this, so that you could bookmark PURLs 14:38:08 TBL: Flight of fancy on 303x, 303y, 303z. . . 14:38:15 "304622 min -- All nobody RESO INVA Adding a live bookmark via feedview uses the location of the feed rather than the location given in the referring page's link element; redirects, PURLs don't work " 14:38:27 maybe this is the bug https://bugzilla.mozilla.org/show_bug.cgi?id=304622 14:38:54 masinter has joined #tagmem 14:39:56 proposed ACTION: Jonathan to research reasons why browser providers (e.g. Mozilla) aren't willing to meet requests (e.g. from purl) to switch address bar URL following successful redirect 14:40:12 ACTION: Jonathan to research reasons why browser providers (e.g. Mozilla) aren't willing to meet requests (e.g. from purl) to switch address bar URL following successful redirect 14:40:12 Created ACTION-348 - Research reasons why browser providers (e.g. Mozilla) aren't willing to meet requests (e.g. from purl) to switch address bar URL following successful redirect [on Jonathan Rees - due 2009-12-16]. 14:40:58 or to *not* switch 14:42:30 JR: 3) Rhys Lewis was working on a finding wrt httpRange-14, but that work stopped when the SWEO note Cool URIs for the SemWeb was published 14:42:51 JR: I think that work should be picked up and made into a finding 14:43:22 ... which would replace/elaborate the email message which currently stands as the resolution of httpRange-14 14:44:14 JR: That was the context for ISSUE-57 at its inception 14:45:09 JR: Additional points that have been added, are my points 4--6 14:46:28 timbl has joined #tagmem 14:46:31 JR: Latest news: AWWSW task force has reported: http://www.w3.org/2001/tag/awwsw/http-semantics-report-20091204.html 14:47:10 JR: A number of forms for this work, of which I'm the main editor 14:47:23 ... helped along by our discussion at the last f2f 14:47:44 ... A lot of text to introduce one key definition: 14:48:55 ... for the phrase "corresponds to", which comes from the definition of the 200 response code, in 2616 and HTTPbis 14:49:53 LM: I wouldn't take this too seriously -- we didn't when we wrote it 14:50:32 JR: We agree entirely. It's the practice which matters to actually pin this down 14:51:04 LM: I note that this story works/should work pretty much for ftp: as well 14:51:22 masinter` has joined #tagmem 14:52:00 JR: Wrt WebArch, 'representation' corresponds to 'entity' or 'content entity' 14:52:47 ... and 'represents' corresponds to 'corresponds to' 14:53:04 LMM: the HTML spec uses 'resource' for what HTTP calls entity. I filed a bug; we'll see... 14:53:29 LM: Note that the correspondence as at a particular instant 14:53:47 JR: Yes, at a particular time 14:53:56 LM: And in a particular context 14:54:46 JR: It's hard to pare things down to the point where we could focus 14:55:02 ... So there's now a bunch of stuff which has been moved _off_ the table 14:56:05 JR: Section *HTTP Exchanges* summarizes what we all know about GET requests 14:58:47 DC: hmm... in pt 5, "preferably"? the server decides which resource the name refers to... 14:58:54 JAR: but an intermediary might get confused 14:59:04 DC: ah... "preferably" makes more sense for intermediaries 15:00:02 TBL: 304? 307? 15:00:22 JR: Yes, step 6 pbly should be clarified wrt responses other than 200 15:00:34 DanC_, you asked to be reminded at this time to wrap up redirections and move to versioned specs 15:00:44 JR: [works through the RDF formalization] 15:01:13 TBL: Why did you avoid 'representation' 15:01:34 JR: Because people objected to giving a URI to something called 'representation' a URI 15:02:50 TBL: All I was concerned is to distinguish the original resource, identified by its URI, and the 'resource' which is some representation of that resource, which also may have a URI, but is not the same 15:02:57 JR: Right 15:04:05 JR: correspondence is a 4-place rel'n between resource, a content entity, a start time and an end time 15:05:16 HST: Context is richer than just time 15:05:35 LM: Accept headers 15:05:46 TBL: But there's still something core 15:05:59 JR: I try to work breadth first 15:07:09 HST: I didn't mean Accept Headers, but rather deixis, e.g. http://localhost/ 15:07:19 DC: or http://my.yahoo.com/ 15:07:40 JR: On to section *What this semantics is careful _not_ to say" 15:07:53 masinter has joined #tagmem 15:08:11 http://tools.ietf.org/html/draft-masinter-dated-uri-05 15:08:54 vs http://tools.ietf.org/html/draft-masinter-dated-uri-06 15:09:51 LM: Server response is a speech act 15:10:48 JR: Precisely -- let's look at some more recent slides 15:11:56 ... How do you prove correctness of an HTTP proxy, cache, API or theory 15:13:07 -> http://lists.w3.org/Archives/Public/www-archive/2009Dec/att-0024/z.html Potatoes don't say anything 15:13:54 bug in "Content negotiation" slide: speaks_for should be corresponds_to 15:14:33 slide21 should have corresponds_to instead of speaks_for in conneg slide (21?) 15:15:09 TOPLAS 1993 ? 15:15:10 (I think of it as BAN logic) 15:15:28 JR: Now make use of Abadi, B??, Lampson and Plotkin logic (ABLP) 15:15:38 ... originally for crypto 15:15:49 ... and access control 15:15:52 masinter has joined #tagmem 15:16:24 (a larch formalization http://www.w3.org/Architecture/iiir-larch/BAN.lsl based on a 1989 SRC Research Report ) 15:16:44 s/B??/Burrows/ 15:21:59 LM: What's good about this is precisely that it qualifies everything with the principal who/which/that says it 15:23:20 JR: Crucial observation -- HTTP defines corresponds_to as follows: 15:23:51 "example.com controls {http://example.com/foo corresponds_to foo}" 15:24:24 s/foo}/E}/ 15:28:34 JR: The domain of "says" is principals, Non-principals don't say anything 15:30:57 JR: Not all resources are principals 15:31:09 NM: Break for 15 minutes 15:43:47 There are two versions of ABLP, the DEC SRC TR from 1991, and the TOPLAS paper from 93 or 94 15:45:13 not to be confused with the earlier BAN paper from 1990, which overlaps in content 15:55:28 NM: Resumed 15:55:55 JR: [Gets to slide 12, reconstruction of httpRange-14] 15:56:05 NM: So this is stronger than the original conclusion? 15:56:09 JR: Yes 15:56:42 ... The original 'resolution' simply constrained the range of the corresponds_to relation 15:56:56 ... but it didn't actually address the original problem 15:58:03 NM: [image conneg example] 15:58:30 JR: This theory as it stands isn't articulated enough to determine the relationship between corresponds_to and speaks_for 15:58:42 NM: Good progress here, wrt httpRange-14 15:59:25 NM: Note that we're OK, mostly, when we ask for, say, the Declaration of Independence, and what we get back has some advertising in a sidebar 15:59:35 ... and I think this can address that 15:59:58 LM: I think this is very good stuff. I hope we can use it to clarify what is meant by Origin 16:00:30 Elaborating the "image conneg example": URI identifies a photo. Conneg used to retrieve either jpeg or gif. They agree up to a point in conveying the photo, but not completely, does the theory allow/explain that? 16:00:36 ... The whole CORS, confused deputy, etc. debate is hampered by a lack of clear definition of precisely this kind of thing: what _is_ an origin, a deputy, etc. 16:00:50 Jonathan's answer was: the theory would need to be elaborated to give a clear answer. 16:00:56 ... Linking SemWeb and Security would be a great thing, possibly a win for both sides 16:01:38 NM: Great idea -- specific action? 16:01:55 DC: I'd like to write this up in a different editorial style 16:02:04 Have we finished JAR's slide set? 16:02:07 JR: Sure 16:02:20 ah 16:02:47 JR: Connects with CAPDESC (?), DARPA-funded xxx 16:02:49 The chair would very much like for Dan to propose an action for himself. 16:02:56 . ACTION Dan write up speaks_for applied to httpRedirections and httpRange using motivating examples 16:03:01 Thank you! 16:03:19 ACTION Dan write up speaks_for applied to httpRedirections and httpRange using motivating examples 16:03:19 Created ACTION-349 - Write up speaks_for applied to httpRedirections and httpRange using motivating examples [on Dan Connolly - due 2009-12-16]. 16:04:42 Pointing out Miller et al's Horton paper: http://www.erights.org/elib/capability/horton/ 16:05:00 re: "delegating responsibility in digital systems" 16:05:12 JAR is babbling about Mark Miller's previous work: DARPAbrowser and CAPdesk (w.r.t our discussion of 307 and what's in the browser URI bar, etc. ) 16:05:30 TBL: Slides done, can we try to find a replacement for 'speaks_for' 16:05:48 TBL: We have a URI, we get a 200 16:06:36 ... Using 'speaks_for' as the relationship which relates content to the resource 16:06:58 ... but if R is a person, the content can't 'speak_for' a person 16:07:15 contexts in which the term gets used "a secure channel from Bob speaks for bob" 16:07:16 ... that is, and entity speaking for the agent 16:07:21 s/and/an/ 16:07:47 q+ how do you get a 200 from a person? 16:07:58 q+ to ask how do you get a 200 from a person? 16:08:14 you get a 200 from a server, where the server speaks for the person 16:08:19 JR: In the old days we sent letters, and my letter did 'speak_for' me 16:09:52 JR: No resource speaks for me, it doesn't say that 16:10:15 (it's clear to me that offline witing is going to be more efficient than group discussion, but if Tim has a clear example, I'm interested to capture it.) 16:11:43 i identifies Pat Hayes 16:12:15 2. 200 from resource identified by i 16:12:43 Slide 9 appears to back Tim 16:12:58 conjecture: 200 response speaks for Pat 16:13:04 HST: Stipulate that we have a URI for Pat Hayes 16:13:46 ... Then your slides appear to say that if I get a ContentEntity from GETting that URI 16:14:15 ... that it a) corresponds_to Pat and therefore, per the 'Controversial Axiom', that it speaks_for Path 16:14:19 s/Path/Pat/ 16:14:26 give us a reason to ask Pat not to assert such things, because it breaks our theory 16:14:50 s/give/JAR: would give/ 16:14:50 JR: Ah -- the ContAx isn't licensed by any existing spec. 16:15:03 ... I think it's useful to explain a lot of WebArch 16:15:21 q? 16:15:28 queue = noah, ashok, Dan 16:15:32 ack noah 16:15:39 TBL: So if it is, we have a reductio wrt Pat saying what he says about that URI 16:15:40 phpht 16:15:44 q+ noah 16:15:46 q- ht 16:15:46 q+ ashok 16:15:47 q+ dan 16:15:53 ack noah 16:16:28 JR: Oh, yes, and, the ContAx should include server says that E speaks for R 16:16:39 ... not E speaks for R directly 16:18:37 noah has joined #tagmem 16:19:03 AM: Looking at R doesn't say any s, then E doesn't (mustn't) say any s 16:19:27 JR: This is meant just to be a restatement of the positive direction 16:19:50 AM: This says E's only role is to say what R says 16:19:57 JR: Yes, that's the ContAx 16:20:02 JAR: yes, advertising conflicts 16:20:13 q+ ht to talk about (re)presentation 16:20:18 q- ashok 16:20:20 ack danc 16:20:23 ack dan 16:20:47 DC: I'm getting useful input, not guaranteed to end up in the same place 16:20:54 LM: Please try to include Origin 16:21:03 DC: Not sure how, but I'll at least try. 16:21:39 HT: I think perhaps there are too many levels at which entities say things. It's clear to me that an XML document says some things, because of the semantics of XML. I.e. the infoset. 16:21:47 TBL: I dispute that it says those things. 16:21:51 DC: I understand both positions. 16:21:53 JAR: Me too. 16:22:34 A potato says "help i'm a potato" ? 16:22:42 HT: I'm being intentionally obtuse in part to get to talking about a 3rd party, which is the interpreter of the message. We often think of this as a human observing a screen, can also be listening to audio. 16:22:52 HT: It's that which ultimately says things. 16:23:05 JAR: Similar to the crypto case, in which the interpreters have to be part of the proof system. 16:23:10 (the dispute between TBL and HT is issue ISSUE-28 fragmentInXML-28; odd that tracker considers it closed when it's plain that the TAG doesn't have consensus.) 16:23:35 TBL: When it's RDF, what it says is the triples it produces 16:23:53 (the resolution in tracker sides with Tim) 16:24:17 HT: Isn't that analagous to my statement that what an XML document "says" is first order the Infoset, and then 2nd order the interpretation of those. 16:24:26 TBL: No, I'm talking about the interpretation of the graph. 16:24:28 HT. Ah. 16:25:20 HT: What I scribed is wrong when I attributed to TBL "what it says is the triples it produces"; should have scribed "what it says is what the triples it produces say" 16:25:25 q? 16:25:29 ack ht 16:25:29 ht, you wanted to talk about (re)presentation 16:25:53 NM: good progress here, great work JR 16:26:02 ... DC is going to try to restate/elaborate 16:27:06 action-201? 16:27:06 ACTION-201 -- Jonathan Rees to report on status of AWWSW discussions -- due 2009-12-01 -- PENDINGREVIEW 16:27:06 http://www.w3.org/2001/tag/group/track/actions/201 16:27:17 . action-201 due 15 Mar 2010 16:28:02 [procedural discussion] 16:28:13 action-201 due 15 Mar 2010 16:28:13 ACTION-201 Report on status of AWWSW discussions due date now 15 Mar 2010 16:28:47 TBL: I'd like to see some interaction with the Tabulator work 16:29:04 ACTION-116 due 31 Dec 2009 16:29:04 ACTION-116 Align the tabulator internal vocabulary with the vocabulary in the rules http://esw.w3.org/topic/AwwswDboothsRules, getting changes to either as needed. due date now 31 Dec 2009 16:29:07 ACTION-201 Due 2 March 2010 16:29:07 ACTION-201 Report on status of AWWSW discussions due date now 2 March 2010 16:30:04 zakim, agenda? 16:30:04 I see 16 items remaining on the agenda: 16:30:06 6. Metadata Architecture (HTTP Semantics): ISSUE-57 (HttpRedirections-57): The use of HTTP Redirection [from DanC_] 16:30:09 10. HTML 5 review: References to versioned specs (#15 in our HTML 5 review topics) etc. [from DanC_] 16:30:11 17. morning break [from DanC_] 16:30:12 3. Metadata Architecture: ISSUE-62 (UniformAccessToMetadata-62): Uniform Access to Metadata [from DanC_] 16:30:15 4. Web Application Architecture [from DanC_] 16:30:16 5. Metadata Architecture: ISSUE-63: Metadata Architecture for the Web [from DanC_] 16:30:19 7. ISSUE-53 (genericResources-53): Generic resources [from DanC_] 16:30:21 8. HTML 5 review: ISSUE-20 (errorHandling-20): What should specifications say about error handling? [from DanC_] 16:30:25 9. HTML 5 review: ISSUE-33 (mixedUIXMLNamespace-33): Composability for user interface-oriented XML namespaces [from DanC_] 16:30:29 11. HTML 5 review: ISSUE-54 (TagSoupIntegration-54): Tag soup integration [from DanC_] 16:30:31 12. HTML 5 review: misc. [from DanC_] 16:30:32 13. ISSUE-50 (URNsAndRegistries-50): URIs, URNs, "location independent" naming systems and associated registries for naming on the Web [from DanC_] 16:30:36 14. ISSUE-34 (xmlFunctions-34): XML Transformation and composability (e.g., XSLT,XInclude, Encryption) [from DanC_] 16:30:39 15. widget URI Scheme [from Larry via DanC_] 16:30:40 16. Admin (next telcon) [from DanC_] 16:30:41 18. lunch break [from DanC_] 16:30:55 Zakim, close item 17 16:30:55 agendum 17, morning break, closed 16:30:56 I see 15 items remaining on the agenda; the next one is 16:30:57 6. Metadata Architecture (HTTP Semantics): ISSUE-57 (HttpRedirections-57): The use of HTTP Redirection [from DanC_] 16:31:20 LM: Could we have used a Link Header in a 404 response? 16:31:24 JR: Yes 16:31:43 LM: But not a link in the body of 404 document itself? 16:31:46 DC: No 16:32:19 LM: But I like the idea of having links in the body, because you can have lots of them 16:33:43 Zakim, take up item versioned 16:33:43 agendum 10. "HTML 5 review: References to versioned specs (#15 in our HTML 5 review topics) etc." taken up [from DanC_] 16:33:57 Zakim, close item 6 16:33:57 agendum 6, Metadata Architecture (HTTP Semantics): ISSUE-57 (HttpRedirections-57): The use of HTTP Redirection, closed 16:33:59 I see 14 items remaining on the agenda; the next one is 16:34:01 10. HTML 5 review: References to versioned specs (#15 in our HTML 5 review topics) etc. [from DanC_] 16:34:16 Topic: HTML 5 review: References to versioned specs (#15 in our HTML 5 review topics) etc. 16:34:38 http://lists.w3.org/Archives/Public/www-tag/2009Oct/0075.html 16:34:42 agenda + HTML Versioning Change proposal [Larry] 16:35:18 http://lists.w3.org/Archives/Public/www-tag/2009Oct/0075.html 16:35:30 This is in relation to ACTION-303 16:38:09 q+ to talk about problems with >requiring< future proofing 16:40:22 ack next 16:40:22 AM: Doesn't this allow me to just support an earlier version? 16:40:23 noah, you wanted to talk about problems with >requiring< future proofing 16:40:39 HST: The 'earliest appropriate' sentence is meant to rule that out. 16:40:48 q+ to wonder whether it is confusing to combine conformance and referencing behaviour in one statement 16:40:53 ... Maybe that needs to be stronger 16:41:04 NM: I have a long history of interest in this 16:41:23 q+ to ask for a reminder of a specific case we're particularly interested in... it was somewhere in the HTML 5 references, yes? 16:41:26 ... I like this as a goal for many circumstances 16:41:47 ... But there are cases where it doesn't work 16:41:59 ... The XML 1.1 experience is illustrative in this case 16:42:21 q+ 16:42:22 ... So we shouldn't _require_ this kind of future-proofing of references 16:42:40 q+ 16:42:42 ... Specifically in terms of systems which are involved in communication 16:42:44 +1 "should future-proof" is too strong. The simple case of citing a frozen spec is fine in many cases 16:42:49 q? 16:42:53 ack johnk 16:42:53 johnk, you wanted to wonder whether it is confusing to combine conformance and referencing behaviour in one statement 16:43:00 q+ to reply to Noah wrt editions vs. versions 16:43:29 Seeing where you're going, Henry, unless new editions >never< allow for new content, I think my concern stands. 16:43:33 JK: Conformant implementations? Should that be separated from what is referenced? Trying to pack too much in? 16:43:42 Or maybe I'm not guessing right as to what your concern/suggestion will be. 16:43:52 ... How references are written is different from what is a conformant implementation 16:43:54 ack DanC 16:43:54 DanC_, you wanted to ask for a reminder of a specific case we're particularly interested in... it was somewhere in the HTML 5 references, yes? 16:44:13 DC: There was a specific case wrt the HTML 5 16:44:15 think IETF tradition is to make the 'future proofing' more part of general policy than being specific in each draft. A1 references B1. When B2 updates B1, implementations of A1 may or may not follow B2 16:44:28 HT: As it stands, there are only stubs in the HTML 5 references. 16:44:29 DC: No. 16:44:47 q+ to point out that anyone using this language assumes there is a contract with future working groups to maintain the operability of the referencing spec, when developing new versions of the referenced spec. Some WGs make various sorts of commitment, some don't, and you can't generalize. The way a sepc evolves may in fact affect two different refering specifications quite differently. 16:44:47 HT: Last I looked. E.g. following link from content-sniffing you got something that just said content sniffing. 16:44:54 q? 16:45:21 http://dev.w3.org/html5/spec/references.html#references 16:45:33 For those reading the minutes, potato is the current screen name of W3C Director Tim Berners-Lee 16:45:54 q+ jar to consider classes of comforming implementations (conforming to various combinations of specs) 16:45:57 We pause to read HTML 5 references section.... 16:46:02 HT: Ah, it's better than it was. 16:46:27 DC: So if we pushed on any of these, we would pbly find the editor would have a reason 16:47:09 HT: E.g. the text in the references says "[CSS] 16:47:09 Cascading Style Sheets Level 2 Revision 1, B. Bos, T. Çelik, I. Hickson, H. Lie. W3C, April 2009." 16:47:18 But links the undated copy. 16:47:27 HST: So what does it mean for an implementor 16:47:43 q? 16:47:58 HST: Specifically, implementors 5 years from now have to figure out what was meant. We're trying to fix that. 16:48:00 ack potato 16:48:00 potato, you wanted to point out that anyone using this language assumes there is a contract with future working groups to maintain the operability of the referencing spec, when 16:48:04 ... developing new versions of the referenced spec. Some WGs make various sorts of commitment, some don't, and you can't generalize. The way a sepc evolves may in fact affect two 16:48:06 ... different refering specifications quite differently. 16:48:28 TBL: If you propose we use the present and the future -- why not earlier ones? 16:48:59 ... As for the future, that depends on the sort of WG and the sort of spec. 16:49:21 ... If the group doesn't commit to back compatibility, you can't rely on it 16:49:41 Is the distinction between "edition" and "version" important? 16:49:58 ... You might try to negotiate a commit from the WG that they wont' change. .. . 16:50:06 s/commit/commitment/ 16:50:18 q+ to mention that there can be issues with 3rd party specs. 16:50:19 ... Or you might just require people to check 16:50:38 Can distinction between "technical specification" and "applicability statement" be useful? "applicability statement" calls out specific dated versions, while general "technical specification" doesn't? Two documents, one of which updates. 16:50:40 ack masinter 16:50:54 ... So it's not clear that we can go with what you propose 16:51:09 LM: I like the difference between edition and version 16:51:38 ... We use to differentiate between applicability statements and language specs. 16:51:54 ... So you would only have to update the appl. statement 16:52:12 ack ht 16:52:12 ht, you wanted to reply to Noah wrt editions vs. versions 16:52:12 LM: Alternatively, you could have policy outside the doc. altogether 16:52:20 q+ to ask how can we apply henry'd text to the specific issue noted? 16:52:31 s/henry'd/henry's/ 16:54:15 NM: YOu haven't addressed my concern 16:54:23 HT: The response to Noah and Tim is to say "yes, all those criticisms apply to unrestricted blank checks" (leaving aside for a sec refs to older versions), by relying on the W3C Policy for Edtions (stepping gently around XML 1.1/10 5th edition in particular), is precisely because it makes this plausible. 16:54:31 NM: Do new editions allow new content? 16:54:33 ... because it wasn't lack of back-compat that broke the XML 1.1 situation 16:54:34 HT: Yes. 16:55:43 NM: Then I still have a problem. See problems deploying XML 1.0 5th edition. A sometimes inappropriate (depending on the specs) expectation is created that implementations that haven't been updated will support new content sourced by those that have been. 16:55:46 JR: Conformance to a spec. that has a variable in it is intrinsically vague 16:55:56 q? 16:55:58 ack jar 16:55:58 jar, you wanted to consider classes of comforming implementations (conforming to various combinations of specs) 16:56:16 ... So there's a time-sensitivity wrt the answer to "does this conform?" 16:56:34 q+ to elaborate on JR's answer 16:56:50 ack noah 16:56:50 noah, you wanted to mention that there can be issues with 3rd party specs. 16:57:00 NM: TBL mentioned SOAP in passing 16:57:05 [AM leaves] 16:57:16 ... SOAP wasn't sure about supporting XML 1.1 16:58:00 ... It depended on the Infoset, and we weren't sure that even if we went to XML 1.1, the Infoset would have been well-future-proofed enough for it all to hold together 16:58:35 ... So in some ways, my willingness to future-proof my references depends on other specs _also_ being well future-proofed 16:58:36 ack next 16:58:37 johnk, you wanted to ask how can we apply henry'd text to the specific issue noted? 16:59:39 HST: Yes, we have a real case of this with XML 1.0 5e and XML NS 3e 16:59:50 q? 16:59:56 zakim, close the queue 16:59:56 ok, noah, the speaker queue is closed 17:00:34 JK: Addressing dated prose in conjunction with an undated URI is separate from xxxx? 17:01:00 LM: My assumption is that the dated ref. is normative 17:01:10 If dated spec A normatively cites undated spec B, and artifact Z conforms to A - what does that mean? Maybe: (1) it conforms to A(B(t)) for some t, or (2) it conforms to A(B(t)) for all t, or (3) if conforms to A(B(t)) for t >= now 17:01:14 DC: Hidden URIs are less significant 17:01:29 ack ht 17:01:29 ht, you wanted to elaborate on JR's answer 17:02:24 HT: Jonathan attempted to answer John. I agree as far as it goes but want to go further. You're right, I was trying to address two problems: 1) dated vs. undated refs conflict, and BTW some peoples' styles to make the URI explict... 17:03:28 (editorially I like including the full, dated URI in a citation, but I much prefer using the document title as the link text.) 17:03:29 HT: ...there are many variations on that 2) usually, all that people tend to say is by grouping into normative and non-normative. It's rare for the conformance section to clarify what is meant by making a reference normative. 17:04:02 FWIW, Dan, though it's clunky, I tend to feel that making both live links, to the same URI, is the least bad approach. 17:04:23 the normative reference speaks for the spec that refers to it 17:04:37 (oh... and I don't like "available at"; I consider the semantics "identified by", and I leave it implicit) 17:05:21 http://lists.w3.org/Archives/Public/public-html-comments/2009Dec/0002.html 17:05:23 zakim, open the queue 17:05:23 ok, noah, the speaker queue is open 17:05:33 Queue is open only for next steps discussion 17:06:25 DC: I asked the HTML 5 editor to add 'work in progress' to links to documents which identify themselves as work in progress 17:06:40 ... The response was 'busywork' 17:07:28 NM: I don't think this can go further unless my concerns and maybe TBL's are addressed 17:07:36 (aha! found some work I did in this area: http://lists.w3.org/Archives/Public/public-swbp-wg/2005Sep/0136 'formally defining W3C's namespace change policy options w.r.t. recent TAG versioning terminology' ) 17:07:53 JR: I thought restricting to editions was good enough 17:08:19 TBL: I had missed that HST meant to constrain to editions, that satisfies me 17:08:48 What I have in mind is something along the lines of: 17:09:23 I think the short para HT proposed is "too clever by half"; it'll only be an effective communication if it recapitulates critical parts of the edition policy 17:09:48 The TAG believes that this is good practice in many cases, but not in all. We recognize that, particularly in cases where no assurance is given that future editions won't support use of new (I.e. previously invalid) content, the advice given here may be impractical. 17:09:58 also, I want to make it clear that it's not the only "template" we endorse by providing more than one template; e.g. another one for really frozen, dated specs 17:10:08 . ACTION: Henry to revise http://lists.w3.org/Archives/Public/public-html-comments/2009Dec/0002.html based on feedback on www-tag/html-comments, and the feedback from TAG f2f 2009-12-09 discussion 17:10:37 whether in practice the "edition" process as specified and executed is sufficient to protect investment is something I'm not qualified to answer. it sounds as if it would be, as specified, if followed, but haven't checked... 17:10:51 close action-303 17:10:52 ACTION-303 Draft text on writing references closed 17:10:57 close action-304 17:10:57 ACTION-304 Write up issue around normative references to particular versions of specs closed 17:11:36 ACTION: Henry to revise http://lists.w3.org/Archives/Public/www-tag/2009Oct/0075.html based on feedback on www-tag and the feedback from TAG f2f 2009-12-09 discussion 17:11:36 Created ACTION-350 - Revise http://lists.w3.org/Archives/Public/www-tag/2009Oct/0075.html based on feedback on www-tag and the feedback from TAG f2f 2009-12-09 discussion [on Henry S. Thompson - due 2009-12-16]. 17:13:25 http://www.erights.org/elib/capability/horton/ 17:13:32 Miller et. al. 17:17:16 rrsagent, pointer 17:17:16 See http://www.w3.org/2009/12/09-tagmem-irc#T17-17-16 17:31:18 http://pinpoint.microsoft.com/en-US/Dallas 17:43:58 Tim, if you're interested in Microsoft's Dallas, it was introduced at their developer's conference a couple of weeks ago. You can go to the transcript of the keynote at http://www.microsoft.com/presspass/exec/ozzie/2009/11-17pdc.mspx and look for the word "Dallas". The video of the keynote, with demos, is at http://cdn-smooth.ms-studiosmedia.com/presspass/mpeg2/1001009_PDCD1_500k.mpg 17:44:12 You can use the transcript to find the right place in the video. 18:06:57 zakim, agenda? 18:06:57 I see 15 items remaining on the agenda: 18:06:58 10. HTML 5 review: References to versioned specs (#15 in our HTML 5 review topics) etc. [from DanC_] 18:07:01 3. Metadata Architecture: ISSUE-62 (UniformAccessToMetadata-62): Uniform Access to Metadata [from DanC_] 18:07:04 4. Web Application Architecture [from DanC_] 18:07:06 5. Metadata Architecture: ISSUE-63: Metadata Architecture for the Web [from DanC_] 18:07:08 7. ISSUE-53 (genericResources-53): Generic resources [from DanC_] 18:07:10 8. HTML 5 review: ISSUE-20 (errorHandling-20): What should specifications say about error handling? [from DanC_] 18:07:13 9. HTML 5 review: ISSUE-33 (mixedUIXMLNamespace-33): Composability for user interface-oriented XML namespaces [from DanC_] 18:07:16 11. HTML 5 review: ISSUE-54 (TagSoupIntegration-54): Tag soup integration [from DanC_] 18:07:19 12. HTML 5 review: misc. [from DanC_] 18:07:21 13. ISSUE-50 (URNsAndRegistries-50): URIs, URNs, "location independent" naming systems and associated registries for naming on the Web [from DanC_] 18:07:25 14. ISSUE-34 (xmlFunctions-34): XML Transformation and composability (e.g., XSLT,XInclude, Encryption) [from DanC_] 18:07:29 15. widget URI Scheme [from Larry via DanC_] 18:07:30 16. Admin (next telcon) [from DanC_] 18:07:31 18. lunch break [from DanC_] 18:07:32 19. HTML Versioning Change proposal [from Larry via DanC_] 18:12:32 rrsagent, draft minutes 18:12:32 I have made the request to generate http://www.w3.org/2009/12/09-tagmem-minutes.html noah 18:15:58 Ashok has joined #tagmem 18:35:58 rrsagent, make logs public-visible 18:36:04 rrsagent, draft minutes 18:36:04 I have made the request to generate http://www.w3.org/2009/12/09-tagmem-minutes.html ht 18:36:11 http://www.w3.org/DesignIssues/PersitentDomains 18:36:23 Zakim, clear agenda 18:36:23 agenda cleared 18:36:25 scribenick: jar 18:36:35 or rather http://www.w3.org/DesignIssues/PersistentDomains 18:39:44 I believe the TAG asked me to review widget: 18:39:46 I did so 18:39:54 the webapps working group replied 18:40:00 i answered their replies this morning 18:40:39 if the TAG would like to review the correspondence and chime in later, then we don't need to take up meeting time here. If you'd like, I can go over what I think the open issues are. Opinions? 18:41:22 http://www.w3.org/2001/tag/2009/12/08-agenda.html 18:41:41 http://lists.w3.org/Archives/Public/public-webapps/2009OctDec/ 18:42:08 see "Comment on Widget IRI" messages 18:42:10 (still working on the agenda) 18:46:34 topic: ISSUE-50 and persistent domains 18:47:02 close action-311 18:47:03 ACTION-311 Schedule discussion of a persistent domain name policy promotion closed 18:47:49 timbl: Above link is old, but background 18:48:27 ... Argument against using http: URIs as names, is that DNS doesn't socially support you. The domain name is rented, not owned. 18:49:47 ... One proposal, if it's broken, fix it. 18:50:09 .... DNS was controlled by IETF, ICANN, and it being up for rent was assumed a good idea 18:50:19 ... now the dangers are becoming known. 18:50:46 ... All the white house pages disappeared when the administration changed (e.g.) 18:51:02 danc: (asks about how that example bears...) 18:51:28 points to http://larry.masinter.net/9909-twist.pdf again 18:51:43 timbl: Many companies put up things that people would like to find later 18:51:55 danc: There is a third-party business around finding things like that 18:52:17 (I don't see how domains would help in either of the supposedly-motivating cases timbl just gave) 18:52:28 timbl: Anyhow. One way to tackle is to make a new TLD that has different rules 18:52:35 Noah_phone has joined #tagmem 18:53:18 ... You might use it for archivable web pages , under a set of rules 18:53:45 ... concerning transfer of rights to other entities so that pages can continue to stay live 18:54:00 points to http://larry.masinter.net/duri.html and previous version http://tools.ietf.org/html/draft-masinter-dated-uri-05 18:54:04 ... there might be a pot of $ to pay for this 18:54:38 ... Problem is to design a social system, maybe as a DNS play, or by setting up a consortium 18:54:50 points to whitehouse.gov 18:55:10 ... Suggesting that to help make this happen, the TAG could write a finding advocating it 18:55:10 q? 18:55:20 q+ to suggest a workshop 18:55:24 q+ to note that it's not any more broken that it could/should be. New domains are not going to get companies to keep their product manuals online or stop the whitehouse from forgetting the previous president, and there's endowed publication with PLOS and there's archive.org and national libraries 18:55:26 ashok: These would be *unalterable* pages? 18:55:32 timbl: To be determined 18:55:32 q+ 18:56:15 ashok: Can you then sell something in this archive space? 18:56:22 q+ to say that the TLD with persistent assignment seems very appealing, restricting the owner's ability to alter the pages doesn't. Seems best approached as an orthogonal issue. 18:56:35 timbl: What transfers is responsibility - not any right to change 18:56:42 ack next 18:56:42 jar: It's a contract with the public 18:56:43 ht, you wanted to suggest a workshop 18:57:29 ht: There are many design points. We could spend time talking about alternatives... 18:58:14 ... I wonder is for the TAG to host a workshop before we write a finding, to scare up a representation of the interested parties 18:58:45 ... a new TLD is a problem for existing URIs that are supposed to have persistent resolution 18:58:53 ... but might be worth paying the cost 18:59:18 ... Another way to go is to talk ICANN into a process around existing domains & persistence 18:59:22 (ah.. that would be better... a way for any domain to get permanent status, sorta like 5013(c) ) 18:59:59 s/5013(c)/501(c)3/ 19:00:13 ... Can we get theorists, library community, other constituencies together to talk 19:00:58 q? 19:01:07 ht: How about a workshop? 19:01:09 +1 workshop 19:01:34 points out talks from previous 1999 workshop on Internet Scale naming 19:01:48 ack next 19:01:48 Wondering whether cost/logistics would work out for workshop proposal. If so, seems appealing, but not sure whether we can get 19:01:49 DanC_, you wanted to note that it's not any more broken that it could/should be. New domains are not going to get companies to keep their product manuals online or stop the 19:01:52 q? 19:01:53 ... whitehouse from forgetting the previous president, and there's endowed publication with PLOS and there's archive.org and national libraries 19:02:06 danc: Tim's examples didn't motivate a TLD for me... 19:02:06 ___________________/me "All your bits are blong to us" - danc 19:02:19 danc: Giving more visible to best practices is a good idea though 19:02:21 s/_//g 19:02:56 danc: There's a running business that does endowed web publication 19:03:06 ht: I haven't found any reference to DNS insurance 19:03:34 danc: There are journals like PLoS that charge authors because they agree to host the content in perpetuity 19:03:55 ... you pay once, it's there forever 19:04:02 noah: (pokes fun at this) 19:04:27 http://www.arkhold.org/ 19:04:29 danc: The White House doesn't have the URI persistence ethic 19:04:39 ack next 19:04:40 points to "This American Life" story about a cyrogenics firm which promised perpetual freezing: http://thisamericanlife.org/Radio_Episode.aspx?sched=1239 19:05:18 masinter: Points to 1999 workshop "problems URIs don't solve" 19:05:29 Q+ 19:05:30 points to http://larry.masinter.net/9909-twist.pdf again 19:06:22 masinter: Organizations split. They merge. They go out of business. Sub-sites move. Countries disappear. 19:06:39 ... In perpetuity has to be around content, not just names 19:07:08 ... People will look to organizations like archive.org for long-term resolvable names 19:07:16 ./me quickly runs a script to change all the links in all his HTML to point to an internet archive version of the URL just in case 19:07:27 ... Getting a guarantee is not the same thing as getting a credible guarantee 19:07:50 points to http://larry.masinter.net/duri.html and previous version 19:07:50 http://tools.ietf.org/html/draft-masinter-dated-uri-05 19:08:22 lm: would like advice on how to progress with these two projects 19:09:14 ... duri = dated URI, guarantees persistent reference, but resolution may be tricky 19:09:51 I wonder whether "that described by" is one word in Latin 19:09:52 ... still puzzled about this approach 19:10:08 danc: Use cases? 19:10:12 q? 19:10:41 q+ to ask what is the incentive for someone to use duri and if not sufficient incentive, and not all using them, wouldn't we still have the problems described by Tim? 19:10:58 lm: tdb: has an optional date... actually two of them, when the resource was read, and when it was interpreted 19:11:58 danc: I've never seen a situation where the complexity of duri: is required 19:12:20 q+ to mention transparency 19:12:46 ack next 19:12:47 noah, you wanted to say that the TLD with persistent assignment seems very appealing, restricting the owner's ability to alter the pages doesn't. Seems best approached as an 19:12:49 ... orthogonal issue. 19:12:53 danc: The URI scheme space is high price real estate, so better to do as an RDF property if possible 19:13:26 s/if possible/for those who are happy to use RDF/ 19:13:29 noah: Something was said about locking down the content, Tim hesitated 19:13:43 timbl: source code repository with version control 19:13:45 1999 workshop: http://www.isr.uci.edu/events/twist/twist99/ 19:14:07 q+ 19:14:11 noah: What about perpetual ownership of name - should be orthogonal to an obligation to preserve 19:14:40 noah: Preservation of content should be more granular 19:14:42 q+ to push back: why should IBM get "ibm.com" in perpetuity without giving back to the commons/community a persistence promise (e.g. re content of homepage) 19:15:08 ack next 19:15:40 ashok: Who will host all this stuff? Not a private company, which can go away. 19:15:52 timbl: A consortium of libraries. 19:16:03 ht: Replication is the only assurance of permanence 19:16:03 q? 19:16:21 ack next 19:16:21 ht: This is a huge design space. 19:16:22 johnk, you wanted to ask what is the incentive for someone to use duri and if not sufficient incentive, and not all using them, wouldn't we still have the problems described by 19:16:23 q+ who gets "att.com" when AT&T is broken up into baby bells, lucent, etc. 19:16:25 ... Tim? 19:16:31 q+ to ask who gets "att.com" when AT&T is broken up into baby bells, lucent, etc. 19:17:10 q+ to suggest we've reached a point of diminishing returns, provided HT will take an action to follow up on the possibility of a workshop... maybe give a date to the resolution of a quarter-year 19:17:15 johnk: This is a social problem. Not sure we can solve this. All of the institutions and agreements go away. 19:17:58 johnk: Not sure this is web architecture 19:18:09 q? 19:18:10 timbl: We need to kick it from the technical into the social 19:18:25 ack ht 19:18:25 ht, you wanted to mention transparency 19:18:26 points to http://www.lockss.org/lockss/Home 19:18:30 johnk: There's no technical solution here 19:18:37 yes, lockss is great work in this space 19:19:18 Heads up: before Dan goes, I want to remind everyone that we should switch to generic resources within 5+ mins 19:19:31 ht: Footnote: The motivation for things like tdb: and wpn: was transparency, so that you can tell by looking at a URI that it named a non-information-resource (not sure i still believe that) 19:19:35 q? 19:20:16 ht: One component is a board of trustees with the power to wind it all up (e.g. if there were no web, at some future time) 19:20:31 points to http://larry.masinter.net/0603-archiving.pdf for long-term archiving also (and see references) 19:20:59 ht: The digital curation people worry about: Where do the resources come from to carry resources forward (e.g. archaic disks) 19:21:11 timbl: lots of ways for accessibility to fail 19:21:33 ack next 19:21:34 DanC_, you wanted to push back: why should IBM get "ibm.com" in perpetuity without giving back to the commons/community a persistence promise (e.g. re content of homepage) and to 19:21:39 ... suggest we've reached a point of diminishing returns, provided HT will take an action to follow up on the possibility of a workshop... maybe give a date to the resolution of a 19:21:41 ... quarter-year 19:22:06 ht: Aim for June? 19:23:31 suggest phrasing, "perhaps in June" 19:23:56 ACTION Henry to look into a workshop on persistence... perhaps the June 2010 timeframe 19:23:56 Created ACTION-351 - Look into a workshop on persistence... perhaps the June 2010 timeframe [on Henry S. Thompson - due 2009-12-16]. 19:24:20 q? 19:24:27 ack next 19:24:28 masinter, you wanted to ask who gets "att.com" when AT&T is broken up into baby bells, lucent, etc. 19:25:00 lm: recommends references in the long term archiving paper (see above) 19:25:09 ... esp the references 19:25:32 NM: To be clear, I think persistence of name assignment should be attacked (mostly) separately from encouraging providers of content to provide that content in perpetuity and/or to keep it immutable. 19:25:37 ack masinter 19:25:42 s/keep it/make it/ 19:26:12 action-312? 19:26:12 ACTION-312 -- Jonathan Rees to find a path thru the specs that I think contradicts Dan's reading of webarch -- due 2009-12-01 -- PENDINGREVIEW 19:26:12 http://www.w3.org/2001/tag/group/track/actions/312 19:26:44 http://lists.w3.org/Archives/Public/www-tag/2009Dec/0061.html 19:26:48 JAR: The email I sent on Monday was sort of "camouflaged" 19:27:18 JAR: In a sense, some people are trying to say, 'I can prove I need URNs 19:27:30 JAR: I was trying to set that down more rigorously. 19:27:49 JAR: I want to relate it to the formalism I've been building. 19:28:02 close action-312 19:28:02 ACTION-312 Find a path thru the specs that I think contradicts Dan's reading of webarch closed 19:28:42 action-121 due 15 Mar 2010 19:28:42 ACTION-121 HT to draft TAG input to review of draft ARK RFC due date now 15 Mar 2010 19:28:49 action-121 due 2 Mar 2010 19:28:49 ACTION-121 HT to draft TAG input to review of draft ARK RFC due date now 2 Mar 2010 19:29:20 topic: Generic resources 19:29:28 action-33 due 20 Dec 19:29:28 ACTION-33 revise naming challenges story in response to Dec 2008 F2F discussion due date now 20 Dec 19:29:58 Topic: ISSUE-53 (genericResources-53): Generic resources 19:30:56 http://lists.w3.org/Archives/Public/www-tag/2009Nov/0069.html 19:30:58 masinter: I drafted replacement text 19:32:09 ... "how to use conneg" explanation for HTTPbis 19:32:57 http://lists.w3.org/Archives/Public/ietf-http-wg/2009JulSep/0763.html 19:34:49 danc: Don't see any text about how the representations relate to one another 19:36:22 BTW, the "problems" with the tag-weekly.html version of the agenda seem to be due to slow response by W3C servers. The tag-weekly.html version now appears to match the dated version. 19:36:28 http://lists.w3.org/Archives/Public/www-tag/2009Nov/0077.html 19:37:54 masinter: sentence about server's purposes needs to be added. re-open action 19:38:30 Noah_phone has joined #tagmem 19:40:00 danc: This is what the speaks_for slide in the presentation is about... if representations contradict, it's incoherent 19:40:44 danc: How about striking "for its purposes" 19:40:56 lm: "for the purposes of this communication" 19:41:00 +1 19:41:04 +1 19:42:26 noah: -1 19:42:45 s/noah: -1/ / 19:42:57 noah: (making another point about attribution) 19:43:50 ...determining, for the purposes of this communication, which representations... 19:43:54 Note that the supplier of representations (or choices) has the responsibility of determining, for purposes of this communication, which representations might be considered to be the "same". 19:44:22 I don't like "considered to be the same". 19:44:35 how about: considered to give the same information 19:45:37 noah: The spec already says entity corresponds to resource 19:45:53 ... Two representations each have the responsibility to correspond to. 19:46:02 ... so nothing else needs to be said. 19:46:12 change "might be considered 'the same'" to "might be considered to represent the same information' 19:46:15 DanC: That's the bug we're trying to fix. 19:46:45 Not convinced. 19:46:51 noah: Saying "corresponds to" is enough 19:47:12 the proposed text in http://lists.w3.org/Archives/Public/ietf-http-wg/2009JulSep/0763.html uses "represent" 19:47:37 johnk: You're saying two things. Do we want to make the second statement, that the conneg reps have to correspond? 19:48:03 There is already an obligation that each representation correspond. It will tend to be the case that multiple representations of a (an immutable) resource will tend to have interpretations that are in some ways similar, perhaps extremely similar, but the archicture should not rule out, e.g. a B&W gif and a color jpeg of very different resolution. 19:48:32 s/to correspond/to sufficiently resemble one another/ (or something similar) 19:48:33 lm: Different ways to represent "the same information" (quoting lm's email 763) 19:49:42 lm: I infelicitously said "same representations" when I should have said "represent the same information" 19:49:50 noah: There are enough weasel words 19:50:03 ... good that we're talking about representing the same information 19:50:08 I.e. to make me happy 19:50:15 lm: And the server has responsibility. 19:50:32 action-231? 19:50:32 ACTION-231 -- Larry Masinter to draft replacement for \"how to use conneg\" stuff in HTTP spec -- due 2009-11-18 -- OPEN 19:50:32 http://www.w3.org/2001/tag/group/track/actions/231 19:51:19 action-231 due next week 19:51:19 ACTION-231 Draft replacement for \"how to use conneg\" stuff in HTTP spec due date now next week 19:51:21 (consensus around give or represent the same information) 19:51:26 ? 19:53:07 break. 19:53:32 noahm has joined #tagmem 19:55:41 noah has joined #tagmem 19:59:31 noahm has joined #tagmem 20:23:14 DanC_ has joined #tagmem 20:24:01 topic: Web Application Architecture (ACTION-306 etc) 20:24:31 noah: Let's see if we can get organized for a more comprehensive approach, or find a whole that's greater than the sum of the parts 20:24:49 ... The TOC is broader in the topic coverage than it might be 20:24:56 q+ to project the web app product next to the outline, and to suggest (a) invited presentations or other get-togethers and (b) looking at relevant wikipedia pages 20:25:19 q+ jar to ask for / suggest criteria etc 20:25:41 noah: maybe look at the form of our products in this area 20:26:15 ashok: From what we spoke about yesterday, it seemed there were many differences between various people think about web apps 20:26:38 ... I thought: web app = you are working with several communicating components 20:27:00 ... but maybe some people thought it was an app running on a server [with sessions] 20:27:02 q? 20:27:23 q+ to respond to ashok 20:27:26 ack next 20:27:27 DanC_, you wanted to project the web app product next to the outline, and to suggest (a) invited presentations or other get-togethers and (b) looking at relevant wikipedia pages 20:27:29 ... In the first case, authorization etc are big issues. In 2nd case, security issues go away 20:28:11 danc: I was looking at PhoneGap and Native Client [see previous action] 20:28:27 ... Inviting any of those folks to talk to us would be a good thing 20:29:00 ... Let's look at wikipedia pages related to security, web apps, widgets, etc 20:29:23 ... The idea is to inform the developer community; a lot of people end up at wikipedia 20:29:48 ... Maybe contributing to wp might be a way to help 20:29:56 ... (brainstorming) 20:30:02 ack next 20:30:03 jar, you wanted to ask for / suggest criteria etc 20:30:29 JAR: I agree with Ashok's comment about Web applications, and assumed we were talking about the distributed case. 20:30:48 JAR: I assumed it involved The Common Man in the Street (TCMITS). 20:31:38 JAR: Regarding the TOC, it was a brain dump, first developed by the group together, and then refined by me. What I'm missing are criteria. Some sort of structure or philosophy that would guide us. 20:31:56 q+ to say that it was part of web arch in 1990 20:32:50 NM muses: maybe the criteria include: 1) architectural issues you would not get right based on what's been set out for the Web of documents and 2) clarifying points of confusion Goal: show that it's, in the end, one consistent, scalable architecture integrating documents and apps. 20:33:44 JAR: Consider, e.g., why a specific programming language wasn't chosen for the Web. It was deemed desirable to have competition there. Maybe there's a winner now (Javascript.) Anyway, what do we want to make the same, and what different? 20:33:54 q? 20:34:48 noah: We don't talk about how you use oracle, that's an implementation detail 20:34:54 (I dunno how conscious it was that javascript happened when it happened... there was talk of active content back in 1990. tcl and such. not to mention display postscript.) 20:35:23 well, and you have XSLT with XML and CSS too I guess 20:35:29 noah: Things like cross-origin security, how to use URIs right - those things are in scope 20:35:40 ... What happens inside server is not in scope 20:36:09 ... typed possible criteria into IRC (above) 20:36:44 ... clarify confusion around e.g. AJAX, or say how to apply old story in new situations 20:37:34 ... to what extent is google maps one application, vs. a very large number of maps? ... more than just a document 20:37:40 q? 20:37:48 q? 20:38:12 ack noahm 20:38:12 noahm, you wanted to respond to ashok 20:39:28 timbl: Even though mapping software allows you to display many overlays, this is always done in code. But with calendars - you can control calendar view, how they're stacked / displayed - that's richer than what you can do with maps 20:39:44 (hmm... I wonder if KML is sufficient.) 20:39:49 I think that talking about proper use of URIs when you're composing layers might be interesting 20:40:01 (... to get maps to work, like calendars, in various clients) 20:40:07 Ah, when Tim says music, he's thinking more iTunes than Sibelius 20:40:36 http://www.sibelius.com/home/index_flash.html 20:40:48 timbl: Music: iTunes maybe - other applications - multidimensional access / view 20:40:54 TBL: Key point is you're looking at more than one document at a time 20:41:00 q? 20:41:01 q? 20:41:48 timbl: When you pull in the data you have to be clever. E.g. you're looking for photos tagged x. Client would do a query to get the photos of interest 20:41:58 (it's really a drag that the Zakim queue isn't a UI feature, e.g. integrated with the list of names in the channel. So many times I'm this close >< to writing an ajax-based front end to Zakim/tracker/rrsagent) 20:42:00 [?] 20:42:25 ack next 20:42:26 johnk, you wanted to say that it was part of web arch in 1990 20:42:31 q+ to ask about use of core mechanisms like URIs in the Tim use case 20:42:36 ack johnk 20:42:54 johnk: Want to push back on jar's idea that webarch didn't address programming / application layer 20:44:00 ... For last TAG meeting I tried to draw a parallel between local web browser vs. javascript ... original web arch did deal with this... 20:44:14 timbl: For example, you could have faceted browsing using forms 20:44:34 ... javascript model just moves data/code onto the client 20:45:35 q? 20:45:38 johnk: Phone's IP address isn't public, but a server [once it knows address] can call back to the phone to perform actions 20:46:11 ... would like to address that applications are distributed in some way [holds up piece of paper] 20:46:39 (timbl takes photo) 20:47:35 johnk: Here are some models. 1. server & client, server assembles a widget, client GETs widget, does a software install 20:47:48 ... interesting thing is 2 trust decisions. 1. Install? 2. Run? 20:48:07 ... side case: What is difference between this and native client, or plugin? 20:48:09 q+ to look at the list of install-time capabilities/permissions in the W3C widgets spec 20:48:31 ... again you have 2 trust decisions, except that (maybe) app is given more power 20:48:58 ashok: Model: app stays on server --- 20:49:47 johnk: I'm not done. Case 2. For example, in iGoogle (?), Google says all this content is sanctioned by Google 20:50:37 ... Client does a GET, trust decision is: Install + run? (as one decision) 20:50:45 ... ashok: How different from widget case? 20:50:52 s/... // 20:51:12 johnk: Both in one step. 20:51:42 noah: (something about cookies vs. user ids) 20:51:50 q+ to note http://www.w3.org/TR/widgets/#feature seems to have no starter-set of permissions. untested hooks are evil 20:52:33 noah: Reserve the word "install" for ... 20:53:27 johnk: Case 3: Site A has a document, with content that calls out to site B (Fedex and airline) 20:54:13 ... Fedex has document that calls out to airline 20:54:41 ... (2nd example) Amazon is in control, compiles the content 20:55:24 ... Cross-site case. there are trust decisions in both directions 20:56:35 danc: Line from amazon to fedex - ? 20:57:01 johnk: Not saying this is deployed in a reasonable way, just observing 20:57:33 Case 4: Client accesses both Amazon and Fedex 20:57:54 ... the client does the mashup 20:58:01 danc: e.g. tabulator 20:58:13 ... We're trying to get a feel for case 4 20:58:42 timbl: Tabulator is a browser extension 20:58:50 danc: What's a good example? 20:59:08 timbl: If you look up me, it pulls up information from wikipedia 20:59:17 danc: No, where the *user* chose both sites? 20:59:38 timbl: What people have we seen? 21:00:21 danc: The interesting difference is that in case 4, the user chooses the sources to be combined. It's not one server referring the user to another. 21:01:31 timbl: Consider two people on twitter, each with a bunch of tweets. 21:02:15 (might have been nice if tim had drawn a separate thingy rather than erasing 4. oh well.) 21:02:21 ... Storage of the data is separate from the... 21:02:39 ... Suppose tweets are to be readable by my friends 21:02:51 ... when someone pulls in tweets, it's because they're in the group 21:03:09 ... tabulator code is completely trusted by C. Runs with user's identity 21:03:57 (hmm... this speaks_for exercise might be an interesting way to look closely at OpenID phishing risks... and to explore my intuition that OAuth is sorta kerberos-shaped) 21:04:52 johnk: The user has to decide to download the twitter app, and ...? 21:04:59 timbl: No, it's in the cloud 21:05:15 (scribe not quite getting it) 21:06:08 timbl: Separate decisions about where to store their data, vs. [something about the app] 21:06:43 johnk: (End of 4 cases as diagrammed on piece of paper and then on the whiteboard) 21:07:12 (jar takes a photo) 21:07:22 (er.. timbl takes a photo) 21:07:47 johnk: web server provider / consumer issues coming out of SOAP work 21:08:09 ashok: There are several trust decisions... made by the *user* explicitly 21:08:25 johnk: brainstorming... 21:08:45 johnk: The site is also making some decisions for you 21:09:01 (jar did not take a photo because tim did) 21:09:46 q? 21:10:08 . ACTION: John integrate whiteboard drawings into a prose document about ways to distribute applications 21:10:15 ashok: In case 2, where igoogle pulls in stuff for you, there's the question of state 21:10:20 johnk: Yes, in all 4 cases 21:10:24 ACTION: John integrate whiteboard drawings into a prose document about ways to distribute applications 21:10:24 Created ACTION-352 - Integrate whiteboard drawings into a prose document about ways to distribute applications [on John Kemp - due 2009-12-16]. 21:10:27 q? 21:10:30 ack noahm 21:10:30 noahm, you wanted to ask about use of core mechanisms like URIs in the Tim use case 21:11:11 noah: Tim's use case was about making maps much better. You go out and say 'tell me about this area' 21:11:18 ack dan 21:11:18 DanC_, you wanted to look at the list of install-time capabilities/permissions in the W3C widgets spec and to note http://www.w3.org/TR/widgets/#feature seems to have no 21:11:21 ... starter-set of permissions. untested hooks are evil 21:11:26 q+ to ask about use of core mechanisms like URIs in the Tim use case 21:11:27 (timbl recessing himself) 21:11:50 danc: List of install capabilities in widget spec - seems dangerous to standardize this 21:12:14 danc: "This is xxx and it wants to look at your contacts list" 21:12:20 (timbl back) 21:12:59 danc: Can't find an actual starter list of particular permissions / capabilities - seems good to not standardize, but seems bad because not tested 21:13:36 ... Lets you sprinkle open dust on your distributed system 21:13:55 noah: We're no worse off. Let the market deal with it 21:14:08 johnk: Symbian has a specific list of caps that the OS gives you 21:14:37 I'm fairly satisfied with using URI space as a marketplace of features, if it works out that way 21:14:48 masinter: Issue of versioning APIs, registries comes up repeatedly 21:15:04 ... the problem becomes much worse regarding what might be available on the device 21:15:20 but yeah... if everybody pretends to support hundred-pound-gorrila.com/featurex , then that sucks 21:15:21 ... "are you a Symbian phone"? is the wrong question. "do you support geolocation?" 21:16:04 noah: If you have an ordinary web page, it asks, can I call the geoloc API? 21:16:26 ... or, in the install process, the question gets asked at install time 21:17:01 ... phonegap either does or doesn't give you a good answer 21:17:28 danc: The premise of the w3c widget spec is that you could have a w3c widget store 21:17:52 noah: The 100-pound gorilla phenomenon is still a risk 21:18:02 s/noah:/danc:/ 21:18:04 q? 21:18:16 danc: ... little guys will be disenfranchised 21:18:24 s/noah: the 100/danc: the 100/ 21:19:00 ack next 21:19:01 noahm, you wanted to ask about use of core mechanisms like URIs in the Tim use case 21:19:11 lm: If you want to name it with the name of the implementation, it's hard to extend, or you run into trademark problems 21:19:15 q+ to ask about use of core mechanisms like URIs in the Tim use case 21:19:22 danc: It's in CR (widget packaging & config) 21:20:07 noah: If they want to write a great iphone app this is a dumb way to do it 21:20:59 lm: The failure hasn't happened because the 2 years haven't passed (you name a capability by the implementation, and there's no extensibility story, then within 2 years you'll have kludges) 21:21:10 jar: +1 to LM 21:21:27 I'm not convinced we're seeing that problem is happening. Yet. 21:21:30 danc: The spec says, URIs go here 21:21:55 I'm sympathetic to watching for this trouble happening; I'm unenthusiastic about getting the TAG all geared up about this until we see trouble brewing. 21:21:57 danc: The install time ritual says, this app wants to look at x, y, z 21:22:02 +1 to Noah 21:22:05 ... The spec only says put URIs here 21:22:34 ... Maybe there will be a marketplace... but maybe the gorilla gets in there, and everyone else has to pretend to be the gorilla 21:22:48 noah: It's not the user-agent string case 21:22:53 danc: No, not interestingly different 21:23:09 I'm not convinced it's underspecified. 21:23:35 lm: If there's part of a spec that's underspecified, and that part need specifications for interoperability, we (TAG) could say so 21:23:36 I think the basis for the widget spec is exactly _for_ interoperability 21:24:00 timbl: Expecting that probably , there will be the equivalent of a mime type registry 21:24:01 I think there will be much more diversity here than for mime types. 21:24:16 q+ to talk about innovation vs. standardizatoin in this space 21:24:34 timbl: current frame, focal length, lots of profiles to talk about... w3c may get involved 21:25:24 noah: The tough thing is there's lots of innovation going on... would have been bad for standardization to rule out multitouch 21:25:32 ... the fact that it's a URI is good 21:26:25 q? 21:26:27 (given that the players in this space seem to be acting in good faith, I'm ok to accept the 100-pound-gorrilla name-mangling risk; I'm OK to hope for a healthy market) 21:26:32 lm: I don't want a solution, I just want to ask the question? What is the migration path e.g. from one pointer to two? 21:26:44 s/question?/question:/ 21:27:06 danc: Maybe people will come to W3C to get a URI? 21:27:06 q+ to ask about use of core mechanisms like URIs in the Tim use case 21:27:34 lm: We'd like to see, if they have a solution, let's get it documented better. If not, let's work on one. 21:28:03 Larry, if you want an action, you can pretty much always assign yourself one. or you can nominate somebody. 21:28:46 ack next 21:28:48 noahm, you wanted to ask about use of core mechanisms like URIs in the Tim use case and to talk about innovation vs. standardizatoin in this space and to ask about use of core 21:28:50 ... mechanisms like URIs in the Tim use case 21:29:19 lm: Not sure I want to engage widget folks again 21:30:27 noah: The maps could be more sophisticated... (that's what Tim was saying...) telling a story about naming and identity is import. Is there agreement on when to mint a URI, how much client/server AJAX flexibility is, who knows what the URIs are. Very interesting area to work. 21:31:02 q? 21:31:17 q+ jar This is the meaning story, not the identity story 21:31:33 q+ to say that for that class of application (map, iTune, document mgt, iPhoto, calendar, timelines, etc) there typically are *not* URIs for the total view. Tabulator has plyed with that. But maybe a language for describingthem is bettre. 21:31:40 noah: TAG story: identity, interaction, formats 21:31:56 danc: Identity per noah is a big story 21:32:30 (scribe hears "semantics" when noah says "identity") 21:32:42 noah: Portals ... 21:33:18 DanC: it's interesting to me in that it includes/subsumes the concern I have about "proposal to make ajax crawlable". If success can be less than the whole thing, I'm all for it. 21:34:58 q? 21:35:33 q+ jar to talk about the US civil war 21:35:39 ack timbl 21:35:39 timbl, you wanted to say that for that class of application (map, iTune, document mgt, iPhoto, calendar, timelines, etc) there typically are *not* URIs for the total view. 21:35:43 ... Tabulator has plyed with that. But maybe a language for describingthem is bettre. 21:35:55 Are not and should not be, or are not but there should be? 21:36:06 timbl: Noah asked, do people make up URIs for the views? 21:36:26 ... Not in general. 21:36:33 ... If so, the URIs get big. 21:36:47 ... Tabulator students took a sparql query to encode a view. 21:37:16 ... When URIs get too big, they invent a data format. 21:37:31 q+ jar to talk about sparql-over-GET + tinyurl 21:37:48 (jar promises to be brief) 21:37:56 q? 21:38:08 ack next 21:38:09 jar, you wanted to talk about the US civil war and to talk about sparql-over-GET + tinyurl 21:38:39 JAR: In nearly every part of this discussion, I see us dancing around, meaning, inference, and contracts. 21:39:14 JAR: Want to encourage people to look at OWL, which is the W3C technology in the inference space (and it's very nice) 21:39:30 q? 21:39:36 Ashok has joined #tagmem 21:40:32 DC: there's a consortium of URL shortening companies 21:41:57 noah: I said, identification is something we could work on 21:42:18 jar: 'Identification' is meaningless without meaning / inference 21:43:39 (discussion of agenda) 21:43:54 s/we could/we could profitably/ 21:45:33 scribe: re OWL, e.g. a specification induces a class of conforming entities. that's DL. one of many possible applications. 21:47:18 . ACTION: Noah to do just a bit of work framing some issues around identification for Ajax apps (remembering the merged maps use case) Due 20 January 2009 21:47:34 johnk: Approach of taking 3 pillars of webarch is good 21:47:42 s/taking/starting with/ 21:47:48 ACTION: Noah to do just a bit of work framing some issues around identification for Ajax apps (remembering the merged maps use case) Due 20 January 2009 21:47:48 Created ACTION-353 - Do just a bit of work framing some issues around identification for Ajax apps (remembering the merged maps use case) Due 20 January 2009 [on Noah Mendelsohn - due 2009-12-16]. 21:49:06 q? 21:52:12 jar: spec / interface naming /v ersioning is one good focus, security is another 21:52:55 danc: Minions, please check client side storage design and look for architectural issues 21:53:04 ashok: web databases? 21:53:06 danc: yes 21:53:50 . ACTION ashok review client side storage apis (web simple storage etc.), looking for architectural issues or other critical problems... or interesting design features the TAG should know about 21:54:03 ACTION ashok review client side storage apis (web simple storage etc.), looking for architectural issues or other critical problems... or interesting design features the TAG should know about 21:54:03 Created ACTION-354 - Review client side storage apis (web simple storage etc.), looking for architectural issues or other critical problems... or interesting design features the TAG should know about [on Ashok Malhotra - due 2009-12-16]. 21:54:19 johnk: I could try to map AWWW section on interaction to parts of webapps TOC that seem related 21:54:44 noah: Interesting, but how about look at interaction story in webapp & findings, and ask: could I tell the Ajax story? 21:54:58 johnk: Yes, I was trying to be more specific, but that's the idea 21:55:47 . ACTION john to explore the degree to which AWWW and associated findings tell the interaction story for Web Applications 21:57:03 ACTION john to explore the degree to which AWWW and associated findings tell the interaction story for Web Applications due: 2 Feb 2010 21:57:04 Created ACTION-355 - Explore the degree to which AWWW and associated findings tell the interaction story for Web Applications due: 2 Feb 2010 [on John Kemp - due 2009-12-16]. 21:57:44 ACTION-355 = john to explore the degree to which AWWW and associated findings tell the interaction story for Web Applications 21:57:54 ACTION-355: john to explore the degree to which AWWW and associated findings tell the interaction story for Web Applications 21:57:54 ACTION-355 Explore the degree to which AWWW and associated findings tell the interaction story for Web Applications due: 2 Feb 2010 notes added 21:58:05 action-355 due 2 feb 2010 21:58:05 ACTION-355 Explore the degree to which AWWW and associated findings tell the interaction story for Web Applications due: 2 Feb 2010 due date now 2 feb 2010 22:01:41 lm: Do we have an exit strategy for ISSUE-50? 22:02:17 q? 22:05:21 lm: The goal of Henry's action is to close the issue, right? 22:05:24 all: yes 22:05:26 ADJOURNED 22:06:33 masinter has left #tagmem 22:22:54 Noah_phone_ has joined #tagmem 22:23:39 timbl has joined #tagmem 22:30:54 noah has joined #tagmem 22:30:54 noahm has joined #tagmem