15:09:04 RRSAgent has joined #xproc 15:09:04 logging to http://www.w3.org/2006/04/20-xproc-irc 15:09:08 Meeting: XML Processing Model WG 15:09:08 Scribe: Norm 15:09:08 ScribeNick: Norm 15:09:08 Date: 20 Apr 2006 15:09:08 Chair: Norm 15:09:09 Agenda: http://www.w3.org/XML/XProc/2006/04/20-agenda.html 15:09:12 Present: Henry, Michael, Andrew, Mohamed, Murray, Alessandro, Norman, Alex, Richard 15:09:23 Topic: Accept this agenda? 15:09:23 -> http://www.w3.org/XML/XProc/2006/04/20-agenda.html 15:09:29 Accepted 15:09:33 Topic: Accept minutes from the previous teleconference? 15:09:33 -> http://www.w3.org/XML/XProc/2006/04/13-minutes.html 15:10:43 s/Accepted/Alessandro suggests 5 before 3, otherwise accepted 15:10:54 Accepted 15:11:09 Topic: Next meeting: 27 Apr telcon 15:11:09 Any regrets? 15:11:11 None given 15:11:27 ACTION: Henry to provide registration page for August f2f 15:12:01 ACTION: Murray to provide local arrangements info for August 15:12:46 MSM: One prominent way to get to the meeting will be to drive. Can we add some questions about car pooling to the registration form? 15:13:09 Murray: I'm thinking about that, I'll see what makes the most sense. 15:13:43 Henry: Let's us a wiki for that instead 15:13:54 who is going to Xtreme markup ? 15:14:24 Topic: Issue 3117: Should parallel execution of step be allowed by the language? 15:14:31 -> http://www.w3.org/Bugs/Public/show_bug.cgi?id=3117 15:14:43 -richard 15:15:01 +??P59 15:15:08 zakim, ? is richard 15:15:08 +richard; got it 15:15:15 Alessandro: This was raised in a call a few weeks ago. 15:15:43 ...I don't know if we need to spend a whole lot of time on it. We probably don't want to add constructs to the language to control this if we can avoid it. 15:17:18 Richard: I hope most of this falls out naturally. If we don't specify the order of execution where it isn't inevitable. That implicitly allows parallel execution. We don't initially have to say much about it. 15:17:52 Richard: If you have two things that could be executed in parallel, maybe they will be. If you want to synchronize them, you have to provide some mechanism, such as reading a document that one is writing. 15:18:13 Alex: I think we shouldn't disallow parallel execution. 15:18:29 Richard: We shouldn't put anything in the language to accidentally prevent it. 15:19:25 Norm: It sounds like we view parallel exec. just as an optimization. 15:20:17 Richard: I take the normal unix pipeline as a model. If you have two processes running, nothing expresses the order except that if one is reading and one is writing, you can be sure the reader will block waiting for the writer. 15:20:39 Richard: Another aspect is that any kind of streaming implies a certain kind of parallelism. 15:22:12 Norm repeats summary. 15:22:51 Richard: Not just features of the language, but also the way we describe the language. A processing model might have unintended consequences that prevented parallelism, we want to avoid that 15:23:49 Richard: An example: we might say that the processing language as if it executed the components in top-to-bottom, left-to-right order which would be bad because it would imply that side-effects (if there are any) occur in a particular order. 15:24:36 Topic: Issue 3118: Should an implementation of the language be allowed to perform caching? 15:24:43 -> http://www.w3.org/Bugs/Public/show_bug.cgi?id=3118 15:26:06 Alessandro: This is a specific question about a particular example. 15:26:19 ...The stylesheet executed by the second step is executed by the first step. 15:26:38 ...Should the pipeline engine be allowed to cache the stylesheet produced by the first step across invocations 15:27:03 ...Can the engine be smart enough to determine that the output will be the same and reuse a cached value. 15:27:37 Norm: I think that what an engine does is not our problem. 15:28:07 Richard: The answer, in some sense, is obviously yes. If the engine can determine that the same results will be produced, then it can use the cached copy. 15:28:15 http://www.w3.org/2002/09/wbs/38398/XProcFTF2/ is now listed as an open questionnaire for our group 15:28:37 ht thanks 15:28:41 Richard: What does it mean for it to be exactly the same? Vanilla XSLT 1.0 stylesheets can't produce any side effects. 15:29:56 ...but care must still be taken to assure that side effects don't happen 15:30:11 Richard: We may need a way to allow authors to express that some components are side-effect free 15:30:14 q+ to endorse the idea of annotation 15:30:29 Alex: It would be interesting to consider annotating the steps 15:30:58 Alex: You may be able to say "never cache" but maybe a smart impl could cache or not as it saw fit otherwise. 15:31:43 Richard: There are some even simpler cases of caching. In the MT pipeline, we compile schemas and cache them. That means the same schema used in two places can reuse the cached copy. 15:32:22 Alex: The more interesting case is where it's produced by the pipeline. 15:33:08 alexmilowski, cache hints like expires in Cocoon ? 15:33:45 Alex: The concept of a dynamicly generated schema isn't far fetched, but URIs that change everytime you read them could be problematic. 15:34:42 HT: The http expires case isn't good enough. The MT engine checks using the http refresh if stale everytime anyone touches a cached resource because there's no way to count on pipeline time and internet time being similar. 15:35:12 HT: The actual time between two uses of a cached object may be wildly different from what you think they are. The only safe thing to do is ask the server each time. 15:35:41 HT: That works on a filesystem too 15:36:09 HT: I'm not sure how that works in the context of documents generated by the pipeline 15:36:11 ack ht 15:36:11 ht, you wanted to endorse the idea of annotation 15:36:51 HT: I think that for practical reasons, I'd be very unhappy to see any requirement of no side-effects imposed on components. 15:37:24 ...I think "escape to program execution" and "synchronous SOAP exchange" are examples of components that cannot have intrinsic gaurantees of no side-effects. 15:37:46 ...There are also cases of components that do database updates. Those components have a side-effect. 15:38:29 Alex: Those aren't (necessarily) examples of pipeline steps communicating through side-effects 15:38:44 Alex: If you're going to have that synchronization problem, you'd setup a dependency for that. 15:39:08 HT: I'm in favor of an approach which has a default and allows the component to assert the opposite. 15:39:30 HT: 1. Not side-effect free; even though my inputs are the same, you can't be sure I'll produce the same output and 15:39:37 [this seems to be an "i am not a function" declaration?] 15:39:52 HT: 2. An expression of out-of-band dependencies. 15:40:20 Norm observes that we've wandered into the issue of side-effects 15:42:21 Norm: I think everyone will agree that if the pipeline knows the output will be the same, it can cache the result 15:42:43 Alessandro: I'm not sure if this is too strong a statement. Consider the case of reading a stylesheet from a URI. 15:44:01 MSM: Side-effects and caching are not seperable questions 15:44:16 Alex: Caching is a feature of the implementation not the language 15:44:37 Richard: Just a big switch will probably be too coarse grained. 15:45:21 Richard: I imagine descriptions for each component type and the XSLT component might, for example, say that it has no side-effects by default. But then on a particular case, you could override it 15:46:06 Alex: The thing that concerns me about being able to say a component has side-effect is that it isn't clear what that means. 15:46:14 Alex: Does it really effect the pipeline running? 15:46:18 zakim, who is on the call? 15:46:18 On the phone I see Ht, Michael, Andrewf, MoZ, Murray_Maloney, Alessandro_Vernet, Norm, Alex_Milowski, richard 15:46:25 zakim, who is making noise? 15:46:35 ht, listening for 10 seconds I heard sound from the following: richard (83%) 15:46:45 Alex: Unless there's some dependency in the flow-graph, what can the processor do. 15:47:15 PGrosso has left #xproc 15:47:48 Alex: Unless we do something like what XSLT does with the document() function, I'm not sure there's a great answer here. 15:48:33 Norm: I wouldn't stream past a component that had side-effects 15:51:19 Norm describes the case of a SOAP service 15:52:22 Alex: You have a pipeline with five steps, each has an auxinput that calls this SOAP service. 15:52:42 Alex: If you cache, you'll get one answer. If you don't cache, you'll get five answers. 15:54:34 Norm will take the question to email 15:55:10 Topic: Any other business? 15:55:15 None. 15:55:20 Adjourned. 15:55:23 -richard 15:55:24 -Murray_Maloney 15:55:24 -Norm 15:55:25 -Andrewf 15:55:25 -Alex_Milowski 15:55:28 -Alessandro_Vernet 15:55:29 -MoZ 15:55:33 -Michael 15:55:36 \quit 15:55:58 -Ht 15:55:59 XML_PMWG()11:00AM has ended 15:56:00 Attendees were Ht, Michael, MoZ, Andrewf, Murray_Maloney, Alessandro_Vernet, Norm, Alex_Milowski, richard 16:02:56 alexmilowski has left #xproc 16:38:49 Norm you forgot the minutes ? 16:39:07 rrsagent, set logs world-visible 16:39:13 rrsagent, draft minutes 16:39:13 I have made the request to generate http://www.w3.org/2006/04/20-xproc-minutes.html Norm 16:39:14 :) 16:39:17 thank you 16:39:26 I'll edit them and move them to the right place now 16:40:07 Norm, ready for next telcon :) 16:40:31 :-) 16:40:36 I have to craft the agenda first 16:42:08 :-P 16:47:10 ht still here 16:47:19 ht ? 16:47:54 you was speaking of putting information in a wiki for F2F ? 16:50:05 Norm, you remember that ? 16:50:24 Yes. I'm sure we can do that when the time is closer. 16:50:27 Are you coming to the f2f? 16:50:37 i hope to 16:50:44 Cool 16:50:57 i hope to see you in france before ;) 16:50:58 Minutes sent: http://www.w3.org/XML/XProc/2006/04/20-minutes.html 16:51:04 at saclay 16:51:10 Right. I've got those tickets already. That's a done deal. 16:51:45 http://norman.walsh.name/2006/itinerary/06-19-xsl 16:53:06 Norm, nice trip :) 16:53:13 :-) 16:53:29 My folks live in England so any trip to Europe usually involves a couple of days with Mom and Dad 16:53:49 you're english ? 16:54:02 Norm, but my mom is 16:54:28 definie s/Norm/No/ 16:54:36 AAUGRRGH 16:54:43 heh 16:57:19 ok because you don't have the beautiful english accent ;) 16:58:24 s/english/british/ 16:59:12 Norm, how do you ask Zakim, when the conference begin ? 16:59:20 i'm on xsl irc 17:00:04 I did it in #xsl 17:00:29 thanks 17:30:36 Norm has joined #xproc 17:43:37 :-) 17:45:26 i'm in :) 17:45:43 not official but sounds good to me 17:46:09 it's Norm(ed) 17:47:40 Actually using it will require more reading than I can afford right now. 17:48:02 rrsagent, bye 17:48:02 I see 2 open action items saved in http://www.w3.org/2006/04/20-xproc-actions.rdf : 17:48:02 ACTION: Henry to provide registration page for August f2f [1] 17:48:02 recorded in http://www.w3.org/2006/04/20-xproc-irc#T15-11-27 17:48:02 ACTION: Murray to provide local arrangements info for August [2] 17:48:02 recorded in http://www.w3.org/2006/04/20-xproc-irc#T15-12-01