02:24:41 LeeF has joined #sparql 07:26:14 ivan has joined #sparql 08:51:18 AndyS has joined #sparql 13:59:25 RRSAgent has joined #sparql 13:59:25 logging to http://www.w3.org/2010/06/30-sparql-irc 13:59:38 Zakim has joined #sparql 13:59:54 zakim, room for 20 13:59:54 I don't understand 'room for 20', AxelPolleres 14:00:09 Zakim, room for 20? 14:00:11 ok, AxelPolleres; conference Team_(sparql)14:00Z scheduled with code 772775 (SPARQL) for 60 minutes until 1500Z; however, please note that capacity is now overbooked 14:00:51 Team_(sparql)14:00Z has now started 14:00:59 +OlivierCorby 14:01:20 +kasei 14:01:37 MattPerry has joined #sparql 14:02:31 + +1.441.865.aaaa 14:02:33 +AxelPolleres 14:02:41 +MattPerry 14:02:43 Zakim, +1.441.865.aaaa is me 14:02:43 +bglimm; got it 14:02:44 +??P22 14:02:46 Zakim, who is on the phone? 14:02:46 On the phone I see OlivierCorby, kasei, bglimm, AxelPolleres, MattPerry, ??P22 14:02:51 zakim, ??P22 is me 14:02:52 +AndyS; got it 14:03:28 EU bridge seems still to be broken... 14:03:43 UK bridge asks for the code, but ignores the input 14:04:00 Anyone else planning to join us on the phone? 14:04:15 calling now 14:04:59 +pgearon 14:06:15 scribe: Gregory Williams 14:06:22 chair: Axel Polleres 14:07:39 AxelPolleres: want to look at organization of test suite, maybe people who were in DAWG can recap how things worked before 14:08:31 AndyS: we collected test cases in RDF. manifests, one per directory, divided into subject-area directories. 14:08:49 ... manifest is a list of tests, test is a pair of query and result 14:08:59 manifest format proposed for SPARQL 1.1 http://www.w3.org/2009/sparql/docs/tests/README.html 14:09:04 ... query case, action is data coming in, and a query to execute 14:09:13 ... result is a result set or a model, depends on the type of query 14:09:19 ... tests are typed into syntax and eval tests 14:09:32 ... syntax tests are just expected to pass, no results 14:09:48 ... no top-level manifest 14:10:18 ... description of the data doesn't follow what's in service descriptions. just name file for named graphs, name is the same as the file. 14:10:46 +q 14:10:48 q+ to ask about results format 14:10:52 AxelPolleres: no top level manifests? 14:11:14 ... need to check that structure. we can structure as we want, probably want to include existing tests 14:11:26 [] :include works for ARQ 14:11:37 pgearon: couple of different types of graphs can be returned. 14:11:49 ... graphs you asked for with a CONSTRUCT, graph which describes a result set. 14:11:59 AndyS: right, but for tests, there isn't a difference. 14:12:45 AndyS: for SELECT and ASK, we used result set format. When you do a compare, you need to know if it is ordered or not. 14:14:05 bglimm: seen some results in turtle, other cases have xml result formats. Can I choose when to use xml or turtle? 14:14:10 ... is the xml results format seen as a graph? 14:14:35 AndyS: context sensitive. srx file is a sparql xml results format. should be with a SELECT query, otherwise test is wrong. 14:14:45 ... file extension determines type. 14:14:49 srx vs n3 vs ... do we, at the moment only use file extensions? 14:15:11 ... srx can be used unless results are order dependent. 14:15:40 ... srx is not order dependent. 14:15:49 +LeeF 14:16:43 solution-seq/slice-results-01.ttl 14:16:49 Looking for the URL now 14:17:25 several broken links in the new test cases http://www.w3.org/2009/sparql/docs/tests/README.html 14:17:26 AxelPolleres: we don't need new test cases for ordering because we don't change ORDER BY? 14:17:35 AndyS: we only need test cases for new material 14:17:54 is somebody in an aviary? 14:18:02 :) 14:18:04 srx is unordered, good to know! 14:18:49 AxelPolleres: any more history from DAWG? 14:19:15 ... LeeF? 14:19:18 topic: manifest format 14:19:20 q+ 14:19:26 ack bglimm 14:19:27 bglimm, you wanted to ask about results format 14:19:32 q- 14:19:34 no 14:19:58 AndyS: does anybody have feedback on tests from last time? hard bits we can make easier? 14:20:11 q+ 14:20:14 Zakim, who is on the phone? 14:20:14 On the phone I see OlivierCorby, kasei, bglimm, AxelPolleres, MattPerry, AndyS, pgearon, LeeF 14:20:19 There were a couple of errors in the documentation 14:20:24 not big ones 14:20:26 ?? contrary by design? 14:20:38 AndyS: you can't parse all tests in the test suite at the same time. two of them conflict. 14:20:48 -LeeF 14:21:06 I thought those were eval tests? So just can't *run* them all at the same time and pass... 14:21:35 vaguely 14:21:43 optional-filter/expr-5-result-not-simplified.ttl 14:21:46 optional-filter/expr-5-result-simplified.ttl 14:21:47 I think 14:21:53 Zakim, mute me 14:21:54 bglimm should now be muted 14:22:01 pgearon: came across a few little issues. 14:22:11 ... would have to dig up details. not major. 14:22:30 +LeeF 14:22:45 AxelPolleres: maybe we should leave old manifests where they are, or make a sparql-1.0 subsection? 14:23:20 AndyS: don't think anything needs fixing in old test cases, but documentation is a different issue. 14:23:38 ... small preference is a single test suite that can be easily used. copying over sparql 1.0 tests and adding new test cases. 14:23:49 ... reason is to make it easier for implementors. 14:24:00 AxelPolleres: anybody oppose that idea? 14:24:04 +1 14:24:25 AxelPolleres: have to copy over old tests into a new directory, check if we want to restructure. 14:24:42 ... maybe keep them in a seperate sparql-1.0 directory. 14:24:55 tests/SPARQL-1.0/... tests/SPARQL-1.1/... 14:25:00 i was on the queue 14:25:11 http://www.w3.org/2009/sparql/docs/tests/README.html 14:25:16 AxelPolleres: manifest format 14:25:20 q? 14:25:26 ack AndyS 14:25:27 q+ to ask about manifest extension for ent. regimes 14:25:49 Zakim, unmute me 14:25:49 bglimm should no longer be muted 14:26:03 ... two issues. I extended the vocabulary for new things (update). 14:26:27 ... update result success/failure, and what the state of the graph store is after the update. 14:27:02 ... idea was to make action have parameters: query, graph store before evaluation, and graph store after evaluation. 14:27:06 http://www.w3.org/2009/sparql/docs/tests/README.html#updateevaltests 14:27:20 (Observation) slightly different design to query datasets - not that it matters that much 14:27:40 ... under discussion: is this new vocabulary ok? new prefix ut: for update tests vocabulary. 14:28:10 ... is it ok to model the graph store in this way? or should we point to the graph store with a single trig file? 14:28:22 AndyS: any reason to not reuse the vocab in the query tests? 14:28:33 AxelPolleres: no vocabulary in query tests that I'm aware of to do this. 14:28:57 qt:data and qt:graphData 14:29:27 AxelPolleres: would there be subproperties to point to name of graph? 14:29:29 re 18 minutes ago: for anyone still interested in the "ordered result" format, they can look at: http://www.w3.org/2001/sw/DataAccess/tests/data-r2/solution-seq/slice-results-01.ttl 14:29:37 AndyS: the object of qt:graphData is both location of data and the graph name 14:29:50 ... those properties hang off the action node in the graph. no indirection through graph store. 14:30:06 ... all relative URIs so they work on the filesystem, off the web. 14:30:29 ut:namedGraph ut:graph ut:named 14:30:38 ... could use service description modeling, but we already have legacy modeling in place. 14:30:55 AxelPolleres: one argument for keeping existing model, people are used to it. 14:31:08 ... another argument for going with new modeling, align with SD vocab. 14:31:18 no strong feeling here either 14:31:42 ... no preference between the two. current modeling in the new doc just because I didn't find the old modeling. 14:31:56 AndyS: mild preference for using SD vocab. 14:32:28 AxelPolleres: the usage in SD is a bit different because right now the unnamed graph points to a file, not the same as in the SD vocab. 14:33:13 AxelPolleres: opinions? strawpoll? 14:33:15 Advice to the editor 14:34:21 strawpoll,, option 1 data graphData from qt, option 2 ut: .... option3: sd: ... option 4: something different : e.g. TriG 14:34:36 1 14:34:51 2, 1 fine as well for me. 14:35:03 0 14:35:08 still no strong feeling :-) 14:35:17 although TriG is appealing... 14:35:21 netrual - noting that TriG is not as stable going forward as other formats. 14:35:24 I use TriG in my Glitter test suite - but don't think we need to push that on the rest of the world at this moment 14:36:00 AxelPolleres: data/graphData then...? 14:37:01 ACTION: to rewrite README.html using qt:data/qt:graphData in the testcase manifest for updates 14:37:01 Sorry, couldn't find user - to 14:37:14 ACTION: axel to rewrite README.html using qt:data/qt:graphData in the testcase manifest for updates 14:37:14 Created ACTION-267 - Rewrite README.html using qt:data/qt:graphData in the testcase manifest for updates [on Axel Polleres - due 2010-07-07]. 14:37:28 entailments? 14:37:37 AxelPolleres: with vocab fixed, now how to structure manifest files. 14:38:02 ... started collecting test cases. good starting points for aggregates, subqueries, negation, project expressions, and property paths. 14:38:45 close ACTION-192 14:38:45 ACTION-192 Collect test cases for subqueries closed 14:39:00 ... open actions on project expressions. 14:39:13 zakim, who is on the phone? 14:39:13 On the phone I see OlivierCorby, kasei, bglimm, AxelPolleres, MattPerry, AndyS, pgearon, LeeF 14:39:24 ... won't be able to get to them for 2 weeks. 14:39:39 ... do we need these seperately? in many cases they need to be with aggregates. 14:39:40 q? 14:40:14 q- 14:40:38 bglimm: might have different test folders for different entailment regimes. 14:40:48 ... or different vocab for declaring entailment details. 14:41:21 ... something like :resultsRDF and :resultsRDFS 14:41:31 ... different answers for different entailment 14:41:37 sd:entailment ? 14:41:45 [] :resultsEntailment 14:42:12 Zakim, who is talking? 14:42:23 kasei, listening for 10 seconds I heard sound from the following: AxelPolleres (28%) 14:42:28 http://www.w3.org/TR/sparql11-service-description/#id247660 sd:entailRegime ? 14:42:33 AxelPolleres: sd:entailmentRegime an option? 14:42:54 ... you want to assign the property to the results? 14:43:08 bglimm: if you provide different results, you'd have to do that for all the tests. 14:43:26 ... run the test and then choose which result set is appropriate to compare against. 14:43:58 AxelPolleres: I would have thought you'd describe data graph with an entailment regime associated with it. 14:44:13 bglimm: always one entailment regime per test. 14:44:26 I thought we modeled this in the SD specifically to be able to do one regime per graph? 14:44:46 did we? I thought this is still open? 14:45:05 AxelPolleres: not sure if we had agreement on entailment per graph 14:45:26 Reality is that test suite is not comprehensive. Design for 90%. More helpful to cover common parts anyway. 14:45:31 bglimm: answer depends on the regime 14:46:11 AxelPolleres: do you expect one regime for all of the graphs (named and default)? 14:46:20 bglimm: would need to know in advance 14:46:32 AxelPolleres: when you write down the data, give properties to state which entailment regime you want. 14:47:03 qt:data . g1 sd:entailmentRegime RDFS ... 14:47:33 ... could define entailment regime per graph. don't know if we want to allow this. 14:47:59 bglimm: means systems that don't support the regime shouldn't run the test. 14:48:21 ... you can still run the test, but will probably fail if you don't support the regime. 14:48:38 we do this already with :approved... need to check the manifest to know if you should run the test. 14:48:43 In DAWG, the output of tests is an EARL report which is what is submitted in the CR process. 14:49:43 AxelPolleres: add that in the manifest? 14:50:01 There is also the vocab file. 14:50:04 ACTION: Axel to add respective example/description of entailment tests in the README 14:50:05 Created ACTION-268 - Add respective example/description of entailment tests in the README [on Axel Polleres - due 2010-07-07]. 14:50:21 karl has joined #sparql 14:50:48 AxelPolleres: open actions for test cases. 14:51:10 ... some proj expr. covered by aggregate tests. 14:51:30 OlivierCorby: I sent test cases for project expressions. 14:52:29 OlivierCorby: there's an error in the first test case. 14:52:43 AxelPolleres: please put future test cases in manifest format. 14:53:19 close ACTION-196 14:53:19 ACTION-196 Collect test cases for Project Expressions. closed 14:54:00 bglimm: three folders of tests already for my tests. 14:54:12 ... can put them all together and update the manifest to declare entailment regimes. 14:54:45 ACTION: axel to put Aggr test cases in manifest format inside CVS 14:54:46 Created ACTION-269 - Put Aggr test cases in manifest format inside CVS [on Axel Polleres - due 2010-07-07]. 14:55:37 ACTION: Olivier to put subqueries, prop paths, project espressions test cases in manifest format and add to CVS 14:55:37 Created ACTION-270 - Put subqueries, prop paths, project espressions test cases in manifest format and add to CVS [on Olivier Corby - due 2010-07-07]. 14:56:02 close ACTION-191 14:56:02 ACTION-191 Collect test cases for aggegate functions closed 14:56:10 close ACTION-193 14:56:10 ACTION-193 Collect testcases for subqueries closed 14:56:25 close ACTION-195 14:56:25 ACTION-195 Collect test cases for Project Expressions. closed 14:56:34 AndyS: does OlivierCorby have CVS access? 14:56:40 OlivierCorby: no. 14:56:50 close ACTION-198 14:56:50 ACTION-198 Collect test cases for property paths closed 14:56:58 ivan? 14:57:18 AxelPolleres: will check with team contacts about CVS access. 14:57:19 yes? 14:57:37 ivan can you check CVS access for Olivier Corby? 14:57:45 ... and can I give you an action on that? 14:58:02 yes 14:58:05 AndyS: took a lot of reorganization work in DAWG to get things sorted out and in the correct syntax, format. 14:58:21 ... if we can get things in the right format early on, it will make things much easier as we progress 14:58:25 ACTION: ivan to get back to Olivier on CVS access 14:58:25 Sorry, amibiguous username (more than one match) - ivan 14:58:25 Try using a different identifier, such as family name or username (eg. ivan, imikhail) 14:58:46 but what Olivier has to do is to send me or Sandro (or both:-) an ssh public key, together with his W3C account name 14:58:57 OlivierCorby? 14:58:58 ... we've got material in email. should get it into CVS in proper formats. 14:59:13 ... important to get things sorted out now and into CVS. 14:59:18 Yes :) 14:59:20 a LOT 14:59:30 ... LeeF and ericP spent a *lot* of time sorting it out last time. 14:59:30 organizing and reorganizing in particular 14:59:37 AxelPolleres, which subdirectory do you want Olivier to have access to? 15:00:00 ... as a group we need discipline. very important to generate tests, but also important to generate them in a way that doesn't cause extra work. 15:00:14 AxelPolleres: we have a problem if we don't have a person responsible for test cases. 15:01:10 ... standard agendum for weekly telecon to collect test cases, discuss examples, dispatch actions to add to the test suite. 15:01:16 thanks AndyS 15:02:22 LeeF: key thing is to point out early on which test cases we'll discuss on the next week's call 15:02:44 standard agendum should include "test cases" collect and approve new test cases. 15:02:45 ... last time, people didn't look because announced at last minute. 15:02:52 ARQ can execute test manifests using command "arq.qtest [manifest]" -- that will at least validate the form (you may disagree with the results!) 15:02:58 ... should do this, but a hard thing to do well. 15:03:10 I can fix it to be maximum SPARQL 1.1 compatible in a few days if that helps anyone. 15:03:20 AxelPolleres: can collect in agenda tests which have been in email. 15:03:46 ... if we have test case editor, can have them watch mailing list and raise them before telecons. 15:03:46 currently, call your files "???.arq" to get SPARQL 1.1 as it works in 1.0 compatibility mode otherwise. 15:04:16 (means we might have a "[] rdf:type :SPARQL_10") 15:04:35 AxelPolleres: what do we do with areas where we don't have test cases at the moment? 15:04:46 ... trivial one is update. should collect update test cases. 15:05:00 ... could somebody volunteer to write down some update test cases? 15:05:14 I can make a start 15:05:22 ... could be just writing down tests we already have in manifest format. browse through email discussions on update. 15:05:41 ACTION: paul to collect initial update test cases and put them in CVS in manifest format. 15:05:41 Created ACTION-271 - Collect initial update test cases and put them in CVS in manifest format. [on Paul Gearon - due 2010-07-07]. 15:05:48 Zakim, mute me 15:05:48 bglimm should now be muted 15:06:00 AxelPolleres: will anybody else volunteer as well? 15:06:05 I can take a shot at it 15:06:34 ACTION: Matt to collect some Update test cases and put them in CVS 15:06:34 Sorry, couldn't find user - Matt 15:06:56 ACTION: Matthew to collect some Update test cases and put them in CVS 15:06:56 Created ACTION-272 - Collect some Update test cases and put them in CVS [on Matthew Perry - due 2010-07-07]. 15:07:42 AxelPolleres: let us know on the mailing list on status updates 15:08:04 ... other areas: Protocol, federated queries, service description... 15:08:29 q? 15:08:29 ... opinions on any of these? 15:08:33 no 15:08:37 q+ 15:08:38 ack bglimm 15:08:39 bglimm, you wanted to ask about manifest extension for ent. regimes 15:08:42 ack bglimm 15:09:13 Zakim, mute me 15:09:13 bglimm should now be muted 15:09:27 pgearon: is SERVICE applied to a releative URL? 15:09:32 Yes. 15:09:42 AxelPolleres: federated query test cases? 15:09:47 And no - parser converts to absolute. 15:09:48 we are at test cases for fed.queries now 15:09:54 pgearon: thinking how to set up a general test for SERVICE keyword. 15:10:11 AndyS: no relative URLs. everything is resolved to absolute URLs before you get to it. 15:10:26 ... you may want relative URLs, but RDF is all about absolute URLs. 15:10:36 AxelPolleres: might be useful to have test cases for that. 15:10:43 ... what do we need for describing federated query test cases? 15:10:56 ... need graph store and services. 15:11:03 ... all of those have pre- and post-states. 15:11:30 SERVICE <../service2> {} will be "SERVICE ...." which is prob not what you want :-) use BASE (but whole query affected) 15:11:33 AxelPolleres: we have SERVICE in the WHERE part of Update, right? 15:12:12 ... we should be fine if we add to the manifest vocabulary some infrastructure to define the datasets of the services. 15:12:56 how about adding something like qt:service [ qt:seviceName ... ;qt:dada ; qt:dataGraph ... ] 15:13:23 s/sevice/service 15:13:27 not sure how the test eval will actually work... 15:13:35 any objections? 15:14:02 pgearon: worried you've moved the hard part out of the query and into the (manifest?) file. 15:14:24 ... at the moment when we describe locations, it's in the manifest. in the protocol you can override graphs and named graphs. 15:14:42 ... there's no such thing in queries. 15:14:58 q+ 15:15:27 AndyS: w3c could run a test service with the right data behind it. 15:15:28 Oooh, I like that :-) 15:15:49 ... if we want something to test the client side, we may need to provide the server side to test against. 15:16:29 AxelPolleres: that is an option. not sure if we need both sever side and local-equivalent based just on the manifests. 15:16:49 AndyS: not sure how W3C would feel about running conformance tests against non-w3c services. 15:17:15 ... we could keep the things sent over the wire simple. BGP against a small amount of data. 15:17:32 ... could put on sparql.org, but then you're depending on that being there and having the right implementation. 15:17:47 ... execution on server side shouldn't be that complicated since testing client side. 15:17:59 aren't we also depending on the data there? 15:18:01 ... naming-wise, sparql.org is relatively neutral. 15:18:03 for kasei: SPARQL 1.0 can set the FROM and FROM NAMED for a query with the protocol. Because of this we can describe them in the test manifest. However, SERVICE is not configurable in this way, and is completely embedded inside the query. Hence we can't just put it into the manifest. 15:18:19 thanks pgearon 15:19:08 AxelPolleres: can discuss with team contacts about options 15:19:26 q+ 15:19:35 q- 15:19:37 ACTION: Axel to ask team contacts on what we can do towards a central service for SPARQL SERVICE testing... 15:19:38 Created ACTION-273 - Ask team contacts on what we can do towards a central service for SPARQL SERVICE testing... [on Axel Polleres - due 2010-07-07]. 15:19:59 AxelPolleres: not sure if it will work. there has to be a guaranteed service running to answer queries. 15:20:21 ... if it needs to do anything complex (join processing), may be difficult 15:20:23 q? 15:20:54 pgearon: can see a few options. first, just put in http localhost. 15:21:14 ... alternatively, put in example.com which won't work but will show what to do. 15:21:27 ... could put in variable. %X or something. 15:21:44 AxelPolleres: need exta vocabulary to indicate what data we expect to be on those sources. 15:22:24 This reminds me of a regular request I've had for parameterized queries 15:22:55 AxelPolleres: do we have volunteers to write down some test cases? I can do some for federated queries in the manifest. 15:23:42 ACTION: axel to draft some extra bit for federated query tests manifest vocabulary plus example. 15:23:42 Created ACTION-274 - Draft some extra bit for federated query tests manifest vocabulary plus example. [on Axel Polleres - due 2010-07-07]. 15:24:03 volunteers for adding test cases for fed query? 15:24:38 AxelPolleres: anything to do for HTTP Update protocol, service description? 15:24:58 I can think of one or two simple SD tests, but it's open-ended by design, so hard to test. 15:25:09 LeeF: protocol is annoying to test. 15:25:40 ... we didn't have rigorous tests last time. not a good approach. haven't thought about if we can do better this time around. 15:25:47 ... don't know how to automate it. 15:25:54 ... tested response codes. 15:26:12 ... Elias Torres may be able to help reconstruct memory of what we did. 15:26:13 Two kinds: client-side and server-side tests 15:26:36 AxelPolleres: could add to test vocab terms for expected http response codes. 15:27:30 LeeF: testing http response codes not very interesting. will be 200 OK. 15:27:44 AxelPolleres: leave it with the editors. 15:27:53 OK 15:27:53 protocol, leave it with the editors for now... 15:27:59 200 15:28:39 service-description equally unclear, even unclearer... 15:29:11 leave sd out of test cases for now. 15:29:43 AxelPolleres: implementation status? 15:29:51 ... how did people implement tests last time? 15:30:14 q+ 15:30:18 q- 15:30:26 http://www.w3.org/2001/sw/DataAccess/impl-report-ql 15:31:30 Current: post REC results included: http://www.w3.org/2001/sw/DataAccess/tests/implementations 15:32:00 greg arguing to improve the explanation of test cases 15:34:27 http://www.w3.org/2001/sw/DataAccess/tests/r2 15:34:41 http://www.w3.org/2009/sparql/docs/tests/README.html#updateevaltests 15:35:03 Tests README = README and formal def? Split? 15:35:38 AxelPolleres: need to think about the structure of these documents 15:35:44 ... need explanation of manifest format 15:35:48 ... list of test cases 15:35:59 Explanation of manifest format, we need a list of test cases, what else? 15:36:01 overview document/paragraph 15:36:31 AndyS: indroductory material would be great. but infinite blackhole of what we can do to make tests easy to use. 15:36:35 ... tradeoff with time. 15:36:44 and where does more introductory material go to, e.g. such that only "approved" docs need to be passed, etc.? 15:36:53 ... always more explanation or tutorials, etc. that we want to create. 15:36:54 Test suite work is pretty thankless - it's not an easy thing to say, but I think things like great documentation are going to exist if someone steps up and creates them 15:36:58 we're not going to hold up the schedule for it 15:37:21 agree. it's just that there didn't feel like there was *any* introductory text last time. 15:37:54 we probably could have written an OK intro in the time we've been discussing this :-) 15:38:11 AxelPolleres: for the moment, leave structure/tests in place, think about how to improve. 15:38:19 ... not sure what else to do with it. 15:38:35 We had 14 reports last time. 15:38:46 hard to know how to improve because there are so few *new* implementors to ask what they have issues with. 15:38:51 How to run the test suite section in either TestCases or Test Cases structure document? 15:39:18 ACTION: axel to add a dummy section in test cases structure doc 15:39:18 Created ACTION-275 - Add a dummy section in test cases structure doc [on Axel Polleres - due 2010-07-07]. 15:39:36 AxelPolleres: can start from old test cases document. 15:40:08 URL? 15:40:30 http://www.w3.org/2001/sw/DataAccess/tests/r2 15:41:04 LeeF: /tests/r2 was automatically generated, but not sure with what. 15:41:10 ... ericP probably knows. 15:41:35 giving perl a bad name! 15:41:45 ACTION: LeeF to find out whether the scripts for auto-generating http://www.w3.org/2001/sw/DataAccess/tests/r2 are reusable? 15:41:45 Created ACTION-276 - Find out whether the scripts for auto-generating http://www.w3.org/2001/sw/DataAccess/tests/r2 are reusable? [on Lee Feigenbaum - due 2010-07-07]. 15:42:05 i thought i was giving eric a bad name :) 15:42:14 ah, maybe 15:42:16 topic: AOB? 15:42:22 q? 15:42:28 q- 15:42:46 AxelPolleres: thanks for the work 15:43:05 ... start seeing tests in CSV soon hopefully. 15:43:26 adjoirned 15:43:28 -bglimm 15:43:30 thank you 15:43:34 -MattPerry 15:43:35 -pgearon 15:43:37 -OlivierCorby 15:43:39