13:55:50 RRSAgent has joined #sparql 13:55:50 logging to http://www.w3.org/2011/10/04-sparql-irc 13:55:52 RRSAgent, make logs world 13:55:52 Zakim has joined #sparql 13:55:54 Zakim, this will be 77277 13:55:54 ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 5 minutes 13:55:55 Meeting: SPARQL Working Group Teleconference 13:55:55 Date: 04 October 2011 13:55:56 zakim, this will be SPARQL 13:55:56 ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 5 minutes 13:56:03 Chair: AxelPolleres 13:56:26 Agenda: http://lists.w3.org/Archives/Public/public-rdf-dawg/2011OctDec/0010.html 13:56:39 SW_(SPARQL)10:00AM has now started 13:56:45 +??P2 13:56:55 zakim, ??P2 is me 13:56:55 +NickH; got it 13:57:04 zakim, mute me 13:57:04 sorry, NickH, muting is not permitted when only one person is present 13:57:09 heh 13:57:14 ha 13:57:29 cbuilara has joined #sparql 13:57:29 why in the world would that be? 13:57:30 +kasei 13:57:36 zakim, mute me 13:57:36 NickH should now be muted 13:58:10 Regrets: SteveH, Chimezie 13:59:22 +??P4 13:59:31 zakim, ??P4 is me 13:59:31 +cbuilara; got it 13:59:42 + +1.781.899.aaaa 13:59:51 zakim, aaaa is Sandro 13:59:51 +Sandro; got it 13:59:57 +??P19 14:00:04 +Bernard 14:00:07 zakim, ??P19 is me 14:00:07 +AndyS; got it 14:00:10 MattPerry has joined #sparql 14:00:18 kasei - not had a chance yet. 14:00:58 + +1.603.897.aabb 14:00:58 AndyS: ok. i'll try to work through it in more depth 14:01:11 zakim, aabb is me 14:01:11 +MattPerry; got it 14:01:40 + +1.540.841.aacc 14:02:04 Zakim, aacc is me 14:02:07 +pgearon; got it 14:03:25 + +1.617.553.aadd 14:03:29 zakim, aadd is me 14:03:29 +LeeF; got it 14:03:55 +??P29 14:04:16 zakim, who's on the phone? 14:04:16 On the phone I see NickH (muted), kasei, cbuilara, Sandro, AndyS, Bernard, MattPerry, pgearon, LeeF, ??P29 14:04:39 zakim, ??P29 is AxelPolleres 14:04:48 +AxelPolleres; got it 14:05:11 zakim, Bernard is Olivier 14:05:11 +Olivier; got it 14:05:13 bglimm has joined #sparql 14:05:50 http://lists.w3.org/Archives/Public/public-rdf-dawg/2011OctDec/0010.html 14:06:18 I can 14:06:23 scribenick: kasei 14:06:27 topic: admin 14:06:39 + +49.731.502.aaee 14:06:40 PROPOSED: Approve minutes from http://www.w3.org/2009/sparql/meeting/2011-09-27 14:06:56 Zakim, +49.731.502.aaee is me 14:06:56 +bglimm; got it 14:07:00 Zakim, mute me 14:07:00 bglimm should now be muted 14:07:08 RESOLVED: Approve minutes from http://www.w3.org/2009/sparql/meeting/2011-09-27 14:07:36 Next regular meeting: 2011-10-11 @ 15:00 UK / 10:00 EST (scribe: Sandro) 14:07:45 regrets for next week 14:07:48 AxelPolleres has joined #sparql 14:07:53 next week regrets: axel & olivier 14:08:17 topic: liaison with RDF WG 14:08:25 Olivier has joined #sparql 14:08:35 AndyS: not much progress... named graphs discussions continue. 14:08:59 ... been waiting on concensus on discussion on RDF Datasets 14:09:11 ... until entire framework is resolved, wouldn't want to make changes. 14:09:29 ... we're not at a point yet where we have to decide. 14:09:37 LeeF: I don't think the current wording is a problem. 14:09:48 ... hasn't caused problems with implementors. 14:10:07 topic: string functions 14:10:28 LeeF: continuing dicussion from Jeen's comment on motivation for string functions 14:10:36 ... 1- vs. 0-indexed functions 14:10:47 ... trying to decide why we don't have an index function 14:11:01 ... substring-before and substring-after handle same use cases 14:11:13 ... Andy proposed adding substring-before, substring-after, and replace functions. 14:11:21 ... some support for that. 14:11:35 ... questions: some other functions from RIF. should we include them? 14:11:48 ... also from greg regarding unicode. 14:11:56 q+ to ask Axel about fn:string-join 14:12:31 kasei: if we ground these functions in the XSD definitions of substring before and substring after, i looked at the definition and wasn't sure i could implement them 14:12:42 q+ to say without "collation" as with other functions. 14:12:45 ... seemed like it required deep understanding of lots of specs to do the pattern matching properly 14:13:05 AndyS: We could add without the versions with collation 14:13:11 AndyS: i think we'd add them without the collation. we haven't adopted collation forms of other functions in the past. 14:13:37 ... implemented in ~3 lines of java. nothing unusual going on. 14:13:51 fine without collation (ie, standard codepoint collation always)? 14:14:29 AxelPolleres: those functions have two variants. with and without collation. 14:14:38 ... proposal here is to adopt the non-collation versions? 14:14:43 AndyS: yes 14:15:11 ack me 14:15:11 AndyS, you wanted to ask Axel about fn:string-join and to say without "collation" as with other functions. 14:15:13 LeeF: other functions: string-join, some stuff about IRIs 14:15:25 AxelPolleres: IRI to URI 14:15:46 AndyS: why do you want to add str-join? what's the difference over what we've already got? 14:15:54 AxelPolleres: I had overlooked it. 14:16:07 AxelPolleres: str-join isn't the same as concat. adding a separator. 14:16:29 ... we didn't have sequences in RIF. arbitrary number of parameters with a separator. 14:16:38 AndyS: you can write the same thing with concat. 14:16:49 ... arbitrary length list is really fixed by syntax 14:17:22 -cbuilara 14:17:24 LeeF: what about the other functions? 14:17:47 AxelPolleres: should go through them since we have them in RIF. 14:18:47 LeeF: what is the IRI to URI function all about? 14:18:55 AxelPolleres: opeates on strings and converts from URIs to IRIs 14:18:57 ah. fixed. thanks LeeF. 14:19:03 +??P10 14:19:15 And it's not reversible? No uri to iri function in RIF? 14:19:16 cbuilara_ has joined #sparql 14:19:40 AxelPolleres: third function is escape-html-uri 14:19:51 ... comes from xpath/xquery functions. 14:20:06 ... escapes non-printable characters 14:20:20 ... for use in attribute values. 14:20:26 -??P10 14:20:40 LeeF: we didn't set out trying to match RIF functions. haven't heard from anybody asking for these functions. 14:20:44 ... inclined to leave them out. 14:20:46 +??P10 14:20:49 zakim, ??P10 is me 14:20:49 +cbuilara_; got it 14:20:59 AxelPolleres: that's fine. 14:21:24 AndyS: seems troublesome to have one-way functions. 14:21:33 ... string-join isn't the same as the F&O function. 14:21:39 ... does RIF define it to be the same as F&O? 14:22:25 AxelPolleres: defines in terms of string-join by converting args (minus the separator) to a sequence 14:22:43 -cbuilara_ 14:22:46 LeeF: I'm not hearing anybody advocating for it. 14:22:50 PROPOSED: Include STRBEFORE, STRAFTER, and REPLACE in SPARQL 1.1 Query (only the versions without collations) 14:23:07 q+ 14:23:11 ack kasei 14:23:36 kasei: should we discuss whether this works with language literals and datatype literals? 14:24:20 see http://www.w3.org/TR/sparql11-query/#func-starts 14:24:32 AndyS: roughly, if the arguments are the same, you can match them. otherwise you can't. 14:25:11 AndyS: does STR() implicitly on the args. 14:25:17 LeeF: they return strings, though. 14:25:38 AndyS: wold have to think about it. either it's a plain string, or it's the same as the input string. 14:25:58 LeeF: inclination would be to have the result be the same as the input. 14:26:35 and consistency with CONTAINS 14:26:38 +??P10 14:26:44 zakim, ??P10 is me 14:26:44 +cbuilara_; got it 14:27:05 LeeF: what does SUBSTR do? returns language tags? 14:27:13 AndyS: yes, returns same form as principle argument. 14:27:19 LeeF: so should do the same thing as SUBSTR. 14:28:13 +1 14:28:18 +1 14:28:22 (to the proposal above) 14:28:34 RESOLVED: Include STRBEFORE, STRAFTER, and REPLACE in SPARQL 1.1 Query (only the versions without collations) 14:29:03 AndyS: will somebody do the tests? 14:29:04 ACTION AndyS Draft text for STRBEFORE, STRAFTER, and REPLACE 14:29:04 Created ACTION-537 - Draft text for STRBEFORE, STRAFTER, and REPLACE [on Andy Seaborne - due 2011-10-11]. 14:29:32 I can do the tests 14:29:45 ACTION: greg to write tests for STRBEFORE, STRAFTER, and REPLACE 14:29:45 Created ACTION-538 - Write tests for STRBEFORE, STRAFTER, and REPLACE [on Gregory Williams - due 2011-10-11]. 14:30:16 LeeF: this should be enough to reply to Jeen? 14:30:23 AndyS: for the first part. haven't dealt with hash functions yet. 14:30:45 LeeF: based on last week's discussion, we have concensus on keeping string functions 1-indexed. 14:31:11 topic: hash functions 14:31:34 http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JulSep/0314.html 14:31:37 LeeF: Andy listed 3 proposals in email. 14:31:57 http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JulSep/0320.html 14:32:32 ... proposal 5 in that link was based on what xml digital signature group includes. 14:32:48 ... discussion on whether MD5 should be included due to security concerns. 14:32:57 sandro: xml dig sig group does not include MD5 14:33:15 Who actively supports any SHA2-isms? 14:33:30 "who" in what sense? 14:33:39 LeeF: variables on proposals being whether we include MD5, whether we include 3 variants of SHA2 14:33:41 "who" = WG member 14:34:06 ... proposal 5 is what's in the draft now but without SHA-224(?) 14:34:17 +1 SHA2 14:34:29 +q 14:34:48 ack pgearon 14:35:15 pgearon: agree MD5 isn't secure, but run into use cases with legacy mysql code using MD5. convenient to be able to talk to those systems. 14:35:28 ... would like to support inclusion of MD5 14:35:37 Oracle Database supports MD4, MD5 and SHA1 14:36:01 sandro: should include some note about MD5 not being secure w.r.t. cryptography, but included for non-crypto uses. 14:36:28 LeeF: easiest proposal is to remove SHA224. small change, addresses comment, mesh well with existing practice. 14:36:31 http://en.wikipedia.org/wiki/MD5#Security "severe compromised" (for security) 14:36:59 PROPOSED: Remove SHA224 from SPARQL 1.1 Query; add text noting that MD5 shouldnot be used for cryptographic applications 14:37:05 +1 14:37:05 Only 2^24.1 ops -> easy! 14:37:06 +1 14:37:12 +1 14:37:25 +1 14:37:28 RESOLVED: Remove SHA224 from SPARQL 1.1 Query; add text noting that MD5 shouldnot be used for cryptographic applications 14:38:04 ACTION AndyS Remove SHA224 from rq25; respond to JB-7 14:38:04 Created ACTION-539 - Remove SHA224 from rq25; respond to JB-7 [on Andy Seaborne - due 2011-10-11]. 14:38:35 q+ to ask about richard's comment 14:38:48 topic: comments 14:39:03 ack AxelPolleres 14:39:03 AxelPolleres, you wanted to ask about richard's comment 14:39:27 AxelPolleres: what do people think about adding clarifying wording w.r.t. RDF datasets? 14:39:42 LeeF: discussed earlier about waiting on RDF WG's discussion about datasets. 14:39:52 +1 to wait ; happy to improve text .... once 14:39:59 LeeF: hold off on replying to Richard 14:40:14 -NickH 14:40:26 Owner for this comment? 14:40:32 close ACTION-527 14:40:32 ACTION-527 Let us know where the impl report code is in git and to put the generated report file somewhere in our CVS space closed 14:40:48 topic: actions 14:40:52 AndyS, I can take the comment. 14:41:11 Cool - do you want to add to wiki/Comments? 14:41:26 AxelPolleres: would Matt and Andy be able to review document before last call? 14:41:26 sure 14:41:48 AndyS: should be able to take a look 14:41:53 Yes ( how much change?) 14:42:01 ACTION: Matthew to review the overview document after Axel sends a go-ahead 14:42:01 Created ACTION-540 - Review the overview document after Axel sends a go-ahead [on Matthew Perry - due 2011-10-11]. 14:42:06 ACTION: Andy to review the overview document after Axel sends a go-ahead 14:42:06 Created ACTION-541 - Review the overview document after Axel sends a go-ahead [on Andy Seaborne - due 2011-10-11]. 14:43:22 How will be at ISWC? 14:43:28 Zakim, unmute me 14:43:28 bglimm should no longer be muted 14:43:36 will be there 14:43:37 I will 14:43:41 I will be at ISWC 14:43:46 I will 14:43:52 SPARQL tutorial ... 14:44:08 bye 14:44:09 -LeeF 14:44:11 \quit 14:44:13 -AxelPolleres 14:44:15 -MattPerry 14:44:15 -pgearon 14:44:16 -Olivier 14:44:17 -bglimm 14:44:19 -cbuilara_ 14:44:22 -Sandro 14:44:29 -kasei 14:44:34 I'll fix the tests for (removal of) SHA224 14:44:37 -AndyS 14:44:38 SW_(SPARQL)10:00AM has ended 14:44:40 Attendees were NickH, kasei, cbuilara, +1.781.899.aaaa, Sandro, AndyS, +1.603.897.aabb, MattPerry, +1.540.841.aacc, pgearon, +1.617.553.aadd, LeeF, AxelPolleres, Olivier, bglimm, 14:44:42 ... cbuilara_ 14:44:50 rrsagent, make records public 14:45:33 greg, will you do the minutes? 14:46:06 yeah 14:46:29 alright, thnx! 14:47:33 http://www.w3.org/2009/sparql/meeting/2011-10-04 15:20:27 should numeric types be acceptable arguments to STRBEFORE? e.g. STRBEFORE(123, "3") 15:24:21 AndyS? 15:24:44 (in a call) 15:24:52 ok 16:01:07 and we're doing the whole captured sub-pattern thing in REPLACE()? 16:15:20 kasei, I found that REPLACE was std pattern - it's lready in Java and the syntax is perl (to within obscure flags) 16:16:37 The examples in F&O made it look more different that it actually is. 16:16:45 Very like sed. 16:27:34 hrm 16:27:57 any thoughts on the strbefore/after numerics question? are we operating simply at the lexical level? 16:29:49 also, when run against, e.g., "foo"@en, should failing to match and returning "the empty string" actually return ""@en? 16:40:34 pgearon has joined #sparql 16:45:04 Zakim has left #sparql 16:56:17 What does strstarts do? strstarts(123, "12") ==> error. strstarts(str(123), "12") => true 17:07:43 so only operates on xsd:string, and plain literals? 17:53:40 kasei - I think thats the way to go (do you?) -- consistency with other ops important. Same out as primary arg (string being inspected - arg1) 17:53:51 need to do the spec details. We shall see. 18:04:49 sorry. queued up too much. were you agreeing that empty-string-return should pass through language tags? 18:07:37 Yes (with caveat I haven't spec'ed and done the details). Seems in the spirit of SPARQL. 18:07:51 yeah, ok. 18:08:48 (Not sure when I will have time to grunge the details.) 18:51:31 i've just committed new tests for STRBEFORE, STRAFTER, and REPLACE. 18:51:49 probably more needed, but this is a first cut. 18:53:25 pgearon_ has joined #sparql