12:08:57 DIGlogger (n=dig-logg@groups.csail.mit.edu) has joined #dig 12:08:57 topic is: DIG meets Mondays, including 7 Apr http://dig.csail.mit.edu/ 12:08:57 Users on #dig: DIGlogger RalphS oshani lkagal DanC sandro ericP daniel-soton Tristan 12:15:45 lkagal has quit (Read error: 110 (Connection timed out)) 13:01:39 lkagal (n=lkagal@32.97.103.224) has joined #dig 13:10:29 lkagal has quit () 13:12:18 timbl (n=timbl@c-24-147-9-199.hsd1.ma.comcast.net) has joined #dig 13:20:30 lkagal (n=lkagal@32.97.103.224) has joined #dig 13:23:09 lkagal has quit (Client Quit) 13:32:27 lkagal (n=lkagal@32.97.103.224) has joined #dig 14:02:56 djweitzner (n=djweitzn@pool-71-178-143-205.washdc.east.verizon.net) has joined #dig 14:09:20 oshani has quit () 14:09:34 lkagal has quit () 14:12:45 lkagal (n=lkagal@32.97.103.224) has joined #dig 14:15:23 lkagal has quit (Client Quit) 14:16:17 lkagal (n=lkagal@32.97.103.224) has joined #dig 14:25:12 oshani (n=oshani@30-7-60.wireless.csail.mit.edu) has joined #dig 14:54:49 Hi Oshani 14:54:55 Hi 14:55:05 Do you have a min ? 14:55:09 sure 14:55:46 oshani has quit () 14:56:46 oshani (n=oshani@30-7-60.wireless.csail.mit.edu) has joined #dig 14:57:04 lkagal, I thought you were in your office :) 14:57:23 The justification UI is currently hardcoded to only display explanations for X air:compliant/air:notcompliantWith Y. Sometime ago we had discussed adding a conf file to the justification ui so that the user could specify which conclusion she wants to see the explanation. I was wondering if you would have a chance to add this in the next few weeks ? 14:57:53 Sorry, I should have specified that I wanted to chat via IRC :) I'm out of town today. 14:58:25 oh okay, I guess I could do that.. could you give me a deadline please? 14:59:38 Hmm - deadline. What other issues/projects are you working on ? 15:00:17 well, currently I am trying to integrate Joe's access control policy creation into Tabulator 15:00:43 but I guess next week is a reasonable deadline for me :D 15:03:05 Do you end of next week or sometime the week after would be reasonable ? 15:04:06 It would be great if it could be done before you left because one of the TAMI UROPS might require it :) 15:04:19 sure! 15:04:27 I am on it then!! 15:04:42 amy (n=amy@30-7-171.wireless.csail.mit.edu) has joined #dig 15:07:15 Great, thanks Oshani ! 15:10:15 lkagal, Mike is here today 15:10:27 Mike Stunes ? 15:10:40 yes, that's him 15:11:11 Oh, I had asked him to start next week because I was away this week. 15:11:42 yeah, he said he misread your email :) 15:13:03 lkagal has quit () 15:35:55 lkagal (n=lkagal@32.97.103.224) has joined #dig 15:38:37 lkagal has quit (Client Quit) 15:59:43 oshani has quit () 16:10:37 lkagal (n=lkagal@32.97.103.224) has joined #dig 16:18:27 timbl has quit () 16:19:21 lkagal has quit () 16:29:12 lkagal (n=lkagal@32.97.102.144) has joined #dig 16:53:48 lkagal has quit () 17:01:26 oshani (n=oshani@30-7-60.wireless.csail.mit.edu) has joined #dig 17:11:24 rho (n=rho@chello213047112079.11.11.vie.surfer.at) has joined #dig 17:13:29 rho has quit (Remote closed the connection) 17:14:35 timbl (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 17:30:54 timbl, are we having a separate Tabulator meeting today? Or is it combined with the DIG meeting today? 17:34:18 achmed (n=xchat@84.240.12.80) has joined #dig 17:40:53 achmed has quit () 17:41:07 djweitzner has quit () 17:42:01 AndyS (n=AndyS@62-31-59-225.cable.ubr12.azte.blueyonder.co.uk) has joined #dig 18:18:07 presbrey (n=joe@c-68-54-79-208.hsd1.fl.comcast.net) has joined #dig 18:27:35 Zakim (n=rrs-brid@homer.w3.org) has joined #dig 18:28:03 zakim, this is dig 18:28:04 ok, RalphS; that matches DIG_weekly()2:30PM 18:28:10 +MIT531 18:28:36 zakim, mit531 has Ralph 18:28:37 +Ralph; got it 18:28:58 + +1.781.961.aaaa 18:29:09 zakim, aaaa is Karyn 18:29:09 +Karyn; got it 18:30:06 zakim, presbrey is Joe 18:30:06 sorry, presbrey, I do not recognize a party named 'presbrey' 18:30:15 zakim, Oshani just arrived in mit531 18:30:15 +Oshani; got it 18:30:15 oops, well, hi 18:31:01 zakim, ShingShing just arrived in mit531 18:31:01 +ShingShing; got it 18:31:33 zakim, Eric just arrived in mit531 18:31:34 +Eric; got it 18:32:36 zakim, Tim just arrived in mit531 18:32:36 +Tim; got it 18:32:47 Meeting: DIG 18:33:00 Regrets: Lalana 18:33:08 +Sandro 18:33:31 Regrets+ DanC 18:34:45 Agenda: https://lists.csail.mit.edu/mailman/private/diggers/2008-May/001493.html 18:35:14 zakim, Yosi just arrived in mit531 18:35:14 +Yosi; got it 18:35:41 Agenda: https://lists.csail.mit.edu/mailman/private/diggers/2008-May/001504.html 18:35:53 +Andy_Seaborne 18:35:57 yosi_s (n=syosi@30-7-220.wireless.csail.mit.edu) has joined #dig 18:36:42 Karyn_ (n=Karyn@c-24-218-143-230.hsd1.ma.comcast.net) has joined #dig 18:36:48 testing 18:36:52 hi Karyn 18:36:56 Chair: Tim 18:36:57 YAY! 18:37:57 Topic: SPARQL Update grammars 18:38:09 Tim: how tabulator uses sparql ... 18:38:33 ... tabulator does http post to an endpoint with the content of the update 18:38:38 ... the server performs the update 18:40:06 AndyS: SPARQL Update is in our open source sparql toolkit 18:40:19 ... Computas is using it in their content management system 18:40:35 ... separates business logic in application layer and sparql database underneath 18:41:04 ... they implement this in java using front-end templates 18:41:15 ... Virtuoso also supports sparql update 18:42:01 ... there's also a version of Joseki that takes sparql update posts 18:42:22 ... for security reasons there are two endpoints with different security 18:43:15 TimBL: otherwise, the sparql server has to verify that the requestor has write access before performing an update 18:43:46 ... our plan was to not allow POST unless the requestor has write access 18:44:25 Andy: we support POST of large SPARQL queries 18:44:39 ... there's not much that distinguishes documents from services 18:45:12 ... mostly, I expect POSTs to have sparql update requests 18:45:14 http://example.org/my-foaf.rdf?query=ASK {<#me> foaf:knows ?whom} 18:46:10 http://example.org/my-foaf.rdf?query=ASK {<#me> foaf:knows ?whom GRAPH { ?s ?p ?o } } 18:46:47 Eric: at a high level, Andy's update protocol is more parallel to SQL while mine is more parallel to the query language 18:47:19 ... the two can be mapped to each other in the case of updating only one graph 18:47:29 Andy: to what does your WHERE refer? 18:47:58 Eric: after an update you have the same WHERE as you would have for a query thus can contain multiple graphs 18:48:21 Andy: you have a notion of a 'working graph' 18:48:34 Eric: I may have cheated and not create a parallel grammar construct 18:49:01 ... I only accept concrete variables, no variables 18:49:20 ... if you give it an unbound as in a triple construction, you simply don't get the triple 18:49:36 http://jena.hpl.hp.com/~afs/SPARQL-Update.html 18:49:37 Andy: my model has two parts; the first looks like a CONSTRUCT but the template is asserted back into the data 18:49:55 ... INSERT template WHERE pattern 18:50:09 INSERT { ?x :q 123 } WHERE { ?x :p 123 } 18:50:32 ShingShing (n=chatzill@30-5-230.wireless.csail.mit.edu) has joined #dig 18:51:01 Eric: I believe my implementation should parse this 18:51:34 INSERT { ?x :q 123 GRAPH { ?x :q 456 } } WHERE { ?x :p 123 } 18:53:17 Eric: this uses the same model that the query language uses for querying graphs or models or whatever the implementation calls them 18:53:21 does this work? (doesn't for me) INSERT { ?x :q 123 } WHERE { ?x :p 123 } 18:53:25 s/or models/or formulae/ 18:54:22 Andy: I'm worried that we're going too deep without consulting other use cases 18:57:35 ... I'm also anticipating that people will want SOAP-style process-oriented stuff 18:58:22 Tim: when you apply a policy, you apply it to what a person can do not the way they do it 18:58:36 ... a policy on executing a POST is really a policy about changing a particular datum 18:59:10 ... I'd like that people can match javascript implementations by describing their intention in sparql and letting the sparql server evaluate the policy 18:59:37 Eric: write policies as graph patterns? 18:59:48 Tim: perhaps simpler than that, e.g. limit to certain predicates 19:00:37 ... don't see it practical to implement a whole Web Services architecture in between 19:01:16 -Sandro 19:01:19 Andy: don't disagree, but many people are taking an incremental approach currently 19:01:30 zakim, Sandro just arrived in mit531 19:01:30 +Sandro; got it 19:01:38 Sandro: what about locking & transactions? 19:02:11 Andy: I've proposed that atomic transactions SHOULD be implemented 19:02:41 ... I'm doubtful we could get agreement on making this MUST 19:03:47 Sandro: want to review where rdf-diff work left off 19:03:50 our diff is where we write the graph to be updates: 19:03:51 Andy's SPARUL: INSERT INTO {

} 19:03:53 ericP's SPARUL: INSERT { GRAPH {

} } 19:03:55 . 19:04:00 TimBL: cwm patch only depends on a single graph 19:05:07 Sandro: diff followed by patch with an intervening update from elsewhere unlikely to do the right thing 19:06:25 Andy: I recently added INSERT DATA in which the clause only permits triples 19:06:51 -Andy_Seaborne 19:08:11 Topic: Plans for summer 19:08:23 presbrey, where y'at? 19:08:38 not hearing anything... 19:08:43 ok 19:08:44 :) 19:08:46 hey there 19:08:48 presbrey, what state are you in (of the union) 19:09:05 FL 19:09:10 Can you call in? 19:09:21 sure, whats the number? 19:09:31 Zakim, what is the passcode? 19:09:32 the conference code is 3441, timbl 19:09:40 Zakim, what is the phone number? 19:09:41 I don't understand your question, presbrey. 19:09:55 +1.617.761;.6200 19:09:58 AndyS has left #dig 19:10:01 +1.617.761.6200 19:10:44 tel://1/716/6716200 19:10:52 tel://1/716/671/6200 19:11:02 +1.617.761.62รดรถ 19:11:08 ../253/5702 19:11:17 + +1.239.247.aabb 19:11:27 zakim, aabb is Presbrey 19:11:27 +Presbrey; got it 19:11:28 Zakim, +1.239.247.aabb is Joe 19:11:28 sorry, ericP, I do not recognize a party named '+1.239.247.aabb' 19:11:46 oh look, my phone number is kept private. thats nice! 19:12:26 that's a hard-wired policy, Joe :) 19:13:21 BASE 19:13:21 PREFIX foaf: 19:13:21 ASK 19:13:21 WHERE { <#presbrey> foaf:knows ?friend . ?friend foaf:openid <%{REMOTE_USER}> } 19:13:37 Joe: I'm working on an access-control system 19:13:51 BASE 19:13:51 PREFIX foaf: 19:13:51 ASK 19:13:51 FROM 19:13:51 FROM ?friend 19:13:52 WHERE { <#presbrey> foaf:knows ?friend . ?friend foaf:openid <%{REMOTE_USER}> } 19:13:59 Joe: ... using librdf, which doesn't follow referenced documents 19:14:45 Tim: [to Yosi] should we enhance cwm to refer to the Web while executing a query? 19:15:19 Joe: I've been pulling out the FROM clauses and building a separate triple store from them, then executing the query on that 19:15:27 ... I can cache these triple stores 19:16:31 Tim: cwm has magic predicates that it will map into sql for reading, but won't write to sql 19:17:03 what do you think, Eric 19:17:17 after I interrogated you about this before your flight out 19:17:28 Tim: effectively, you tell cwm that a predicate maps to a column in the db 19:17:47 FROM vs subgraphs 19:18:06 Yosi: cwm only knows about memory store; it is not architected to store things on disk 19:18:19 presbrey, reading back now... 19:18:50 http://presbrey.xvm.mit.edu/common.rdf 19:18:57 Tim: cwm is careful to not blindly load all the URIs it finds; only the ones it needs to evaluate a query 19:19:02 http://presbrey.xvm.mit.edu/.metadata.rdf 19:19:29 Joe: I load all URIs to try to satisfy WHERE 19:19:33 ahh yes, the [[FROM ?friend]] and [[<#presbrey> foaf:knows ?friend]] point. i recall now 19:20:48 Joe: I pull in everything mentioned literally in FROM and things referenced from WHERE 19:21:01 the algae impl allows this: 19:21:03 BASE 19:21:04 PREFIX foaf: 19:21:04 ASK 19:21:04 FROM 19:21:06 WHERE { <#presbrey> foaf:knows ?friend . 19:21:08 GRAPH ?friend { ?friend foaf:openid <%{REMOTE_USER}> } } 19:21:18 Joe: I strip out FROM ?friend first 19:21:33 that is, if told to, it uses GRAPH as sufficient instruction to read a graph 19:22:02 but GRAPH can also reference a different graph at the same URI 19:22:21 I like using FROM because, to me at least, it specifically means get these other URIs 19:22:49 may be available 19:23:39 Eric: the default graph is consistent at a service that supports multiple graphs but you get to further constrain the query 19:24:32 Tim: should we promote sparql services doing tabulator-style recursion? 19:25:18 lookup 19:25:38 Eric: without syntactic clues I fear people will object to scale effects 19:26:01 WHERE { <#presbrey> foaf:knows ?friend . 19:26:01 GRAPH ?friend { ?friend foaf:openid <%{REMOTE_USER}> } } 19:26:03 vs. 19:26:06 "if I get it now, I won't have to get it later", right? 19:26:07 WHERE { <#presbrey> foaf:knows ?friend . 19:26:07 ?friend foaf:openid <%{REMOTE_USER}> } 19:26:15 the ignorant FROM prefetching is essentially for Apache performance 19:27:05 FROM NAME FROM NAME ... 19:27:05 WHERE { <#presbrey> foaf:knows ?friend . 19:27:05 GRAPH ?friend { ?friend foaf:openid <%{REMOTE_USER}> } } 19:27:26 ericP, I do like the 1st one best, but librdf's SPARQL implementation doesn't support subgraphs ;-) 19:27:30 Tim: the general follow-your-nose is that you can assume when you've asked for data about a variable you have looked up that variable 19:28:14 ... this is like loading an HTML file that pulls in images and CSS and the CSS pulls in other things 19:28:49 ... [the client] actually wants all those inclusions to be read 19:29:23 Eric: big CSS providers will flatten their CSS 19:29:38 ... in the SemWeb the analog might discourage people from sharing URIs 19:31:03 ... I think query authors will want syntactic control 19:31:51 Yosi: I don't see a -load flag for subjects in cwm, only one for predicates 19:32:02 presbrey, regarding the editing the ACLs, is it possible for me to edit http://presbrey.xvm.mit.edu/.metadata.rdf? 19:32:20 s/FROM NAME FROM NAME /FROM NAME FROM NAME / 19:33:24 Topic: 2 minutes around the table 19:33:36 ShingShing: just getting started with tabulator 19:33:47 ... learning how to read the code 19:33:55 ... don't yet understand the workflow 19:34:15 oshani, /afs/csail.mit.edu/group/dig/rdf-acl/www/.metadata.rdf 19:34:19 Tim: see test.js 19:34:26 Oshani: working on the policy parser 19:35:31 Joe: I moved .metadata.rdf into AFS so Oshani can edit it directly if the server is down 19:35:39 Yosi: writing 19:36:26 not hearing Yosi clearly 19:36:55 presbrey, is the metadata file webdav editable? 19:37:13 no haven't hooked PUT into access control yet 19:37:17 Yosi: I should finish my thesis before August 19:39:38 http://dig.csail.mit.edu/2008/06/TabulatorHowWhy.svg 19:39:50 Tim: ^ a partial dump of my whiteboard 19:40:14 ... I'd like the students to be working on individual projects you can write up but also for everything to fit together into one system 19:44:02 ... green bits on TabulatorHowWhy.svg are parts that may not need work this summer 19:44:38 ... solid lines indicate dependencies 19:44:48 ... dotted lines indicate 'satisfies' 19:46:15 ... one alternative is to spend this summer making tabulator completely solid 19:46:34 ... another alternative would be to complete pub/sub 19:46:48 Eric: CambridgeSemantics is doing a lot in pub/sub 19:47:03 ... based on open source stuff from IBM 19:47:12 s/based on/adding to/ 19:48:13 ... Joe is going to work on the server side of pub/sub, right? 19:48:17 Joe: for data wiki? 19:49:15 Tim: yes, but hopefully we'd add it to many servers 19:49:59 Sandro: I don't mind a blocking HTTP GET 19:50:13 most gmail users don't appear to either 19:50:29 gdocs, etc 19:51:31 Tim: one implementation is an HTTP POST tunneled in the reverse direction through an HTTP GET 19:52:22 Joe: would be nice if sparql updates where returned in the same stream after the initial document 19:53:10 s/ where /were/ 19:53:18 s/updateswereret/updates were ret/ 19:55:07 presbrey has quit ("Leaving") 19:55:23 Sandro: I like blocking GET because this is simple in Web arch; there are lots of failed pub/sub protocols lying around 19:56:04 Tim: can layer things on jabber, which is fairly well deployed 19:56:45 Sandro: we'd like torrents to work for Web pages too 19:57:00 Tim: blocking GET assumes a server 19:57:04 swiftfox again using 102% of my CPU 19:57:10 Sandro: no, just an address 19:57:37 presbrey (n=joe@c-68-54-79-208.hsd1.fl.comcast.net) has joined #dig 19:58:47 ... easier for developers to write clients than servers 20:00:20 Oshani: are tabulator2007 and tabulator2008 different sets of code? 20:00:24 Tim: no, they're milestones 20:01:43 ... send me input on TabulatorHowWhy.svg 20:01:57 Sandro: anyone but Tim going to Linked Data Planet in New York? 20:02:16 ... I'm thinking of going but it's expensive 20:03:22 oshani has quit () 20:03:30 -Presbrey 20:03:42 [adjourned] 20:03:54 http://dev.w3.org/cvsweb/perl/modules/W3C/Annotations/cgibin/annotate?rev=1.196&content-type=text/x-cvsweb-markup 20:03:54 :) 20:03:57 search for this: BookmarkQuery 20:04:11 Eric: on reimplementing Annotea in SPARQL 20:04:33 Note the Algae2 query: 20:04:33 $AnnotationQueryObject::InputBookmarkQuery = "${NS}ask 20:04:33 (?bookmark rdf:type b:Bookmark {\\%ATTRIB == <\$self->{-attribution}>}. 20:04:33 ?bookmark b:hasTopic ?hasTopic. 20:04:33 ?bookmark b:recalls ?recalls) 20:04:36 collect (?bookmark)"; 20:04:44 oshani (n=oshani@30-7-60.wireless.csail.mit.edu) has joined #dig 20:04:44 That same query expressed in SPARQL: 20:04:44 $AnnotationQueryObject::InputBookmarkQuery = "${NS}SELECT (?bookmark)"; 20:04:44 {GRAPH <\$self->{-attribution}> {?bookmark rdf:type b:Bookmark. 20:04:44 ?bookmark b:hasTopic ?hasTopic. 20:04:44 ?bookmark b:recalls ?recalls}} 20:06:23 -MIT531 20:06:36 -Karyn 20:06:37 DIG_weekly()2:30PM has ended 20:06:39 Attendees were Ralph, +1.781.961.aaaa, Karyn, Oshani, ShingShing, Eric, Tim, Sandro, Yosi, Andy_Seaborne, +1.239.247.aabb, Presbrey 20:06:43 zakim, bye 20:06:43 Zakim has left #dig 20:09:05 timbl has quit () 20:11:45 timbl (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 20:12:45 timbl_ (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 20:12:45 timbl has quit (Read error: 104 (Connection reset by peer)) 20:14:03 presbrey has quit ("Leaving") 20:20:22 RalphS has quit ("bye for today") 20:27:14 timbl (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 20:27:14 timbl_ has quit (Read error: 104 (Connection reset by peer)) 20:28:19 timbl has quit (Read error: 104 (Connection reset by peer)) 20:28:42 timbl (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 20:30:06 timbl has quit (Read error: 104 (Connection reset by peer)) 20:30:27 presbrey (n=joe@c-68-54-79-208.hsd1.fl.comcast.net) has joined #dig 20:31:18 timbl (n=timbl@30-7-141.wireless.csail.mit.edu) has joined #dig 20:33:27 timbl has quit (Client Quit) 20:42:10 Karyn__ (n=Karyn@c-24-218-143-230.hsd1.ma.comcast.net) has joined #dig 20:42:10 Karyn_ has quit (Read error: 104 (Connection reset by peer)) 20:53:10 lkagal (n=lkagal@32.97.103.224) has joined #dig 20:53:48 presbrey has quit (Read error: 113 (No route to host)) 20:55:58 lkagal has quit (Client Quit) 20:59:52 lkagal (n=lkagal@32.97.103.224) has joined #dig 21:12:55 ShingShing has quit ("ChatZilla 0.9.82.1 [Firefox 2.0.0.14/2008040413]") 21:30:16 yosi_s has quit (Read error: 110 (Connection timed out)) 21:39:57 lkagal has quit () 21:41:42 oshani has quit () 21:43:52 presbrey (n=joe@c-68-54-79-208.hsd1.fl.comcast.net) has joined #dig 22:03:53 oshani (n=oshani@SYDNEYPACIFIC-ONE-SEVENTY-SEVEN.MIT.EDU) has joined #dig 22:06:00 Karyn (n=Karyn@c-24-218-143-230.hsd1.ma.comcast.net) has joined #dig 22:06:00 Karyn__ has quit (Read error: 104 (Connection reset by peer)) 22:58:42 lkagal (n=lkagal@32.97.103.4) has joined #dig 23:51:52 oshani has quit () 23:55:31 Karyn has quit ()