IRC log of sparql on 2011-12-19

Timestamps are in UTC.

14:58:13 [RRSAgent]
RRSAgent has joined #sparql
14:58:13 [RRSAgent]
logging to http://www.w3.org/2011/12/19-sparql-irc
14:58:15 [trackbot]
RRSAgent, make logs world
14:58:15 [Zakim]
Zakim has joined #sparql
14:58:17 [trackbot]
Zakim, this will be 77277
14:58:17 [Zakim]
ok, trackbot; I see SW_SPARQL(GraphStore HTTP)10:0 scheduled to start in 2 minutes
14:58:18 [trackbot]
Meeting: SPARQL Working Group Teleconference
14:58:18 [trackbot]
Date: 19 December 2011
14:59:07 [sandro]
Zakim, who is on the call?
14:59:07 [Zakim]
SW_SPARQL(GraphStore HTTP)10:0 has not yet started, sandro
14:59:08 [Zakim]
On IRC I see RRSAgent, AndyS, MacTed, AxelPolleres, swh, kasei, iv_an_ru, trackbot, pgearon, NickH, ya, ericP, sandro
14:59:12 [chimezie]
chimezie has joined #sparql
14:59:23 [sandro]
zakim, what is the code?
14:59:23 [Zakim]
the conference code is 77277 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), sandro
14:59:47 [AndyS]
zakim, start meeting
14:59:47 [Zakim]
I don't understand 'start meeting', AndyS
14:59:56 [AndyS]
zakim, this is SPARQL
14:59:56 [Zakim]
ok, AndyS; that matches SW_SPARQL(GraphStore HTTP)10:0
15:00:04 [sandro]
zakim, what is the code?
15:00:04 [Zakim]
the conference code is 77277 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), sandro
15:00:17 [AndyS]
zakim, who is on the phone?
15:00:18 [Zakim]
On the phone I see no one
15:02:16 [AndyS]
zakim, help
15:02:16 [Zakim]
Please refer to http://www.w3.org/2001/12/zakim-irc-bot for more detailed help.
15:02:18 [Zakim]
Some of the commands I know are:
15:02:19 [Zakim]
xxx is yyy - establish yyy as the name of unknown party xxx
15:02:22 [Zakim]
if yyy is 'me' or 'I', your nick is substituted
15:02:24 [Zakim]
xxx may be yyy - establish yyy as possibly the name of unknown party xxx
15:02:27 [Zakim]
I am xxx - establish your nick as the name of unknown party xxx
15:02:29 [Zakim]
xxx holds yyy [, zzz ...] - establish xxx as a group name and yyy, etc. as participants within that group
15:02:31 [Zakim]
xxx also holds yyy - add yyy to the list of participants in group xxx
15:02:34 [Zakim]
who's here? - lists the participants on the phone
15:02:37 [Zakim]
who's muted? - lists the participants who are muted
15:02:39 [sandro]
zakim, who is here?
15:02:39 [Zakim]
mute xxx - mutes party xxx (like pressing 61#)
15:02:41 [Zakim]
unmute xxx - reverses the effect of "mute" and of 61#
15:02:44 [Zakim]
is xxx here? - reports whether a party named like xxx is present
15:02:46 [Zakim]
list conferences - reports the active conferences
15:02:48 [Zakim]
this is xxx - associates this channel with conference xxx
15:02:50 [Zakim]
excuse us - disconnects from the irc channel
15:02:53 [Zakim]
I last learned something new on $Date: 2010/03/15 18:49:04 $
15:02:57 [Zakim]
SW_SPARQL(GraphStore HTTP)10:0 has ended
15:02:59 [AndyS]
zakim, list
15:02:59 [Zakim]
Attendees were
15:03:01 [Zakim]
SW_SPARQL(GraphStore HTTP)10:0 has now started
15:03:03 [Zakim]
On the phone I see no one
15:03:05 [Zakim]
On IRC I see chimezie, Zakim, RRSAgent, AndyS, MacTed, AxelPolleres, swh, kasei, iv_an_ru, trackbot, pgearon, NickH, ya, ericP, sandro
15:03:10 [Zakim]
I see SW_SPARQL(GraphStore HTTP)10:0, XML_SchemaW()10:00AM, WF_OGD()10:00AM, SW_HCLS(Disc)10:00AM, MM_MMI(Arch)10:00AM, Team_(PHILIPP)9:30AM, WAI_PFWG(ARIA)10:00AM active
15:03:13 [Zakim]
also scheduled at this time are INC_WEBID(WEBID)10:00AM, Team_(VALID)9:00AM, Team_SysWeb()9:30AM
15:03:18 [chimezie]
?
15:03:27 [sandro]
zakim, who is on the call?
15:03:27 [Zakim]
On the phone I see no one
15:04:20 [AndyS]
who is here?
15:04:22 [AndyS]
AndyS
15:04:25 [kasei]
Greg
15:04:26 [sandro]
Sandro
15:04:28 [chimezie]
Chime
15:04:44 [sandro]
swh ?
15:04:51 [swh]
I'll have to redial
15:04:54 [swh]
sound problems
15:04:57 [swh]
I can hear you though
15:05:15 [swh]
complicated :)
15:06:39 [AndyS]
scribe: AndyS
15:06:45 [AndyS]
scribenick: AndyS
15:07:15 [AndyS]
Sandro: I think we are hear to talk about my review of the graph store protocol.
15:08:02 [AndyS]
... and I have concerns with the doc being published.
15:08:21 [AndyS]
... the chairs wanted consensus for TC tomorrow.
15:08:59 [AndyS]
... and whole WG approves at TC
15:10:46 [AndyS]
AndyS: I'm OK to publish 2LC as is.
15:10:49 [sandro]
http://lists.w3.org/Archives/Public/public-rdf-dawg/2011OctDec/0398.html
15:11:30 [AndyS]
sandro: On POST to a graph, then it says the server does a merge.
15:11:37 [swh]
q+
15:12:08 [AndyS]
... I think that is a problem because lots of things appear as RDF graphs e.g. a collection of graphs
15:12:24 [AndyS]
... which itself has RDF representation
15:13:13 [ericP]
q+ to say "saying that POSTs to a resource SHOULD append that resource" reduces the applicability to other usage patterns
15:13:15 [AndyS]
chime: would you be language to cover additional things and maybe in SD
15:14:55 [AndyS]
ericP: if no SPARQL endpoint then SPARQL update on a REST-deployment. Certain language suggests don't use POST.
15:15:15 [sandro]
"The group will also define protocol to update RDF graphs using ReSTful methods."
15:15:16 [AndyS]
sandro: concerned applicablity of this document
15:15:23 [AndyS]
q+
15:15:37 [ericP]
q-
15:15:46 [sandro]
ack sandro
15:15:49 [sandro]
ack swh
15:16:41 [AndyS]
swh: We use POST as append and care about this. MAY is not desirable.
15:16:43 [sandro]
ack AndyS
15:17:10 [ericP]
AndyS: what's your disposition of 2626 and its def'n of POST?
15:17:41 [ericP]
sandro: 2626 doesn't restrict POST. i think constraining the semantics of POST should only be applied to certain resources
15:18:41 [sandro]
q?
15:18:58 [ericP]
AndyS: [cites 2616] POST is used for:
15:19:02 [ericP]
... creating a resource
15:19:05 [ericP]
... @@1
15:19:11 [ericP]
... appending a database
15:20:19 [AndyS]
chime: if you target a service, the doc does not say what happens
15:20:36 [swh]
the doc only describes graphs though
15:21:21 [AndyS]
chime: semantics for common case ... could add lang to allow other case
15:21:38 [AndyS]
sandro: I want to target an object on a web
15:23:14 [sandro]
q+ "graph"
15:24:36 [AndyS]
chime: graph container approx RDF content -- "graph container" label is recent
15:25:09 [AndyS]
q?
15:25:14 [AndyS]
ack "graph"
15:26:00 [AndyS]
chime: IRI ids a thing that can emit graphs? What about language to override -- if documented
15:26:46 [AndyS]
ericP: Concrete use case ...
15:27:34 [AndyS]
... there is a resource for trouble tickets - GET is the list of tickets -POST is add to list.
15:27:55 [AndyS]
q+ to ask if there any POST would not allow.
15:28:13 [AndyS]
... this is what IBM uses it for (ptr?)
15:28:26 [AndyS]
... do we want to use this doc to talk about that?
15:28:40 [sandro]
ack AndyS
15:28:40 [Zakim]
AndyS, you wanted to ask if there any POST would not allow.
15:28:46 [ericP]
AndyS
15:29:04 [ericP]
AndyS: in this scheme, is there anything whihc POST would disallow?
15:31:23 [ericP]
POST /hereComesTrouble { [ a :TroubleTicket ; :poster "bob" ; :issue "borken" ] }
15:32:04 [ericP]
-> { <> :myTickets ( <ticket1> <tickiet2> <ticket3> ) . }
15:32:17 [AndyS]
chime: appending to a graph is the mimimal op - I often go outside this for REST-ful APIs
15:32:49 [AndyS]
q+
15:32:58 [ericP]
... <ticket3Holder> { <ticket3> [ a :TroubleTicket ; :poster "bob" ; :issue "borken" ] }
15:33:42 [AndyS]
q-
15:35:07 [swh]
+1 to ericP, it's all out of band now
15:36:46 [AndyS]
+1 to chime: the GSHP defines XYZ - don't want it => don't claim to use it.
15:37:24 [sandro]
eric: "SHOULD" is there so clients can assume/try it
15:37:33 [sandro]
chime: right
15:38:15 [sandro]
eric: So if one usage has a SHOULD, and people assume/try it, then other users are a SHOULD NOT, which means (1) clients will not try it, and (2) there's a social pressure -- you're not being a good netizen.
15:38:20 [AndyS]
q+ to ask about the scope of the spec.
15:38:37 [Zakim]
restarting irc bot in 1 minute to recover bridge state
15:38:45 [sandro]
chime: What about SHOULD do one thing, MAY do another thing?
15:39:11 [sandro]
eric: What about, "There is a subclass of these, which SHOULD have this behavior"
15:39:11 [sandro]
+1
15:39:21 [sandro]
chime: that makes sense
15:39:32 [sandro]
chime: Some description in SD.
15:39:49 [swh]
q+
15:40:06 [kasei]
q+
15:40:18 [sandro]
eric: anchor in document, a class "Appendable", and maybe make it a MUST. So in the SD someone an say this thing is an Appendable, or this thing creates Appendables.
15:40:19 [Zakim]
Zakim has joined #sparql
15:40:29 [sandro]
chime: that sounds reasonable to me. Steve?
15:40:37 [sandro]
q?
15:40:57 [ericP]
AndyS: i think it's clear that we're not an RDF version of REST
15:41:04 [ericP]
... that means POST means append
15:41:17 [sandro]
andy: I think it's clear we're NOT doing an RDF version of rest. We are JUST doing an RDF Graphstore protocol. If you use that, you're buying into these constraints.
15:41:19 [ericP]
... if you want a behavior, then it's not this protocol
15:41:32 [sandro]
andy: If you want a different set of features, then don't claim conformance to this spec.
15:41:35 [sandro]
-1 Andy
15:41:50 [sandro]
(er, not Andy himself. +1 Andy. -1 Andy to the view)
15:42:14 [sandro]
andy: These scenarios keep coming up that are hypothetical.
15:42:16 [sandro]
q?
15:42:19 [sandro]
q+
15:42:33 [sandro]
andy: It seems useful document that "for a graphstore, it means append"
15:42:41 [sandro]
eric: But it comes in a big package deal.
15:43:01 [sandro]
eric: Can this document be repurposed for the trouble ticket scenario, with SPARQL update, date.
15:43:15 [sandro]
chime: If you assume we're not targetting "RDF Graph Content" then yse.
15:44:01 [sandro]
andy: Just because it has a representation as an RDF graph, doesn't mean this document applies to it.
15:44:14 [sandro]
sandro: I so disagree
15:44:28 [sandro]
andy: It seems like you're trying to put other ideas.
15:45:09 [sandro]
"It should be possible to manipulate RDF graphs using HTTP verbs, notably PUT, POST and DELETE. By this, clients doesn't need to know the SPARQL language to update graphs when it is not needed. "
15:45:44 [sandro]
andy: There's a Web Application working group. Isn't this trouble ticket system a Web Application.
15:46:12 [sandro]
(nonsense. by that argument, the SPARQL Protocol could be done by the Web Apps WG.)
15:47:27 [AndyS]
q-
15:49:40 [AndyS]
http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Jun/0000.html
15:52:45 [swh]
q+
15:53:50 [sandro]
q-
15:57:32 [sandro]
ack swh
15:59:13 [sandro]
sandro: So, Steve, your opinion is that this document doesn't apply to other folks doing REST+RDF.
15:59:40 [swh]
Zakim, who's speaking?
15:59:40 [Zakim]
sorry, swh, I don't know what conference this is
15:59:52 [sandro]
zakim, this is sparql
15:59:52 [Zakim]
ok, sandro; that matches SW_SPARQL(GraphStore HTTP)10:0
15:59:52 [AndyS]
zakim, this is 77277
15:59:53 [Zakim]
AndyS, this was already SW_SPARQL(GraphStore HTTP)10:0
15:59:55 [Zakim]
ok, AndyS; that matches SW_SPARQL(GraphStore HTTP)10:0
16:00:03 [sandro]
zakim, who is taking?
16:00:03 [Zakim]
I don't understand your question, sandro.
16:00:53 [swh]
Zakim, who's speaking?
16:01:04 [Zakim]
swh, listening for 10 seconds I heard sound from the following: +1.617.324.aaaa (55%)
16:01:51 [kasei]
q+ to ask how you "self identify" as a member of this class
16:02:02 [swh]
+1 to kasei's q, I don't get that
16:02:16 [sandro]
eric: PROPOSAL: We identify a class of resources for whom POST = graph merge, and for those resources it's a MUST. Then we let people use some out-of-band mechanism to indicate what resources are in that class.
16:02:25 [AndyS]
Is it not that lable "graph container"?
16:03:01 [sandro]
eric: ... because the rest of this document applies to a broader set of resources.
16:04:02 [sandro]
sandro: No, we need to leave POST open.
16:05:07 [sandro]
entend until 0:20
16:05:09 [sandro]
extend
16:05:24 [sandro]
(20 after the hour)
16:06:27 [sandro]
eric: definitons, the commonsense mapping of REST to RDF, with semantics of POST = graph merge.
16:07:15 [sandro]
eric: how about the rest of the document talk about RDF Resoruces in general, then this section say there is a subset which are POST=append resource. And none of that is tied to it being a sparql graphstore.
16:07:46 [sandro]
chime: not all those things are paired with a graph in the dataset of a graph store
16:08:23 [AndyS]
+1 to Chime. Leave for LDP-WG? RDF-WG?
16:10:43 [ericP]
Note that this section applies only to the application of POST to RDF graph content. other ...
16:10:53 [AndyS]
This doc does not apply to "data-handling process" (a phrase from RFC 2616)
16:10:56 [sandro]
chime: This behavior of POST would only apply when the target is RDF graph content [ SOmething that leaves it open for Eric's use case ]
16:11:18 [sandro]
eric: and strengten this to a MOST.
16:11:21 [sandro]
MUST
16:11:40 [AndyS]
i.e. not applying to "Providing a block of data to a data-handling process."
16:12:52 [sandro]
sandro: so "POST MUST be understand as merge, unless this is a data handling process"
16:13:05 [sandro]
andy: "extending a database through an append opperation"
16:13:38 [sandro]
so, POST MUST be understood if this is the kind of resource for which POST mean appending ...
16:14:02 [sandro]
andy: "graph container"
16:14:09 [sandro]
sandro: would need "only"
16:15:59 [kasei]
q?
16:21:03 [Zakim]
- +1.781.899.aabb
16:21:40 [Zakim]
- +1.617.324.aaaa
16:22:18 [kasei]
ack kasei
16:22:18 [Zakim]
kasei, you wanted to ask how you "self identify" as a member of this class
16:25:07 [AndyS]
zakim, who is on the call?
16:25:07 [Zakim]
On the phone I see ??P1, [IPcaller], kasei, +1.216.368.aacc
16:25:27 [AndyS]
zakim, IPcaller is me
16:25:27 [Zakim]
+AndyS; got it
16:27:24 [AndyS]
/rdf-graphs/service?graph=... ==> /graphstore?graph=...
16:28:03 [swh]
I have to go guys. bye
16:28:06 [Zakim]
-??P1
16:28:54 [AndyS]
Suggested: RDF Graph Content ==> RDF Graph Container
16:28:58 [AndyS]
Suggested: RDF Graph Content ==> Graph Container
16:29:32 [AndyS]
(to follow recent RDF-WG terminology)
16:30:10 [Zakim]
- +1.216.368.aacc
16:30:12 [Zakim]
-AndyS
16:30:15 [Zakim]
-kasei
16:30:16 [Zakim]
SW_SPARQL(GraphStore HTTP)10:0 has ended
16:30:18 [Zakim]
Attendees were +1.617.324.aaaa, +1.781.899.aabb, kasei, +1.216.368.aacc, AndyS
16:31:08 [kasei]
I don't think that makes nearly as much sense, sandro.
16:32:42 [sandro]
*shrug* Okay, yeah, graphstore?graph= is okay, I guess. Certainly better than service?graph=
16:37:09 [AndyS]
sandro - it is not a SPARQL endpoint (either kind) - that is, I think, the source of the confusion.
16:37:36 [AndyS]
endpoint is a (XML) web services concept (WSDL)
16:37:52 [AndyS]
and also a more general, non web concept.
16:38:05 [AndyS]
e.g. telecoms
16:52:33 [AndyS]
swh - kasei has noticed that all <cite> cause no link to be generated.
16:52:55 [AndyS]
Even if inside <a href=></a> then <cite> does not reach HTML.
16:53:53 [AndyS]
Propose to de-<cite> the whole doc.
16:54:09 [AndyS]
ericP?
16:54:37 [kasei]
AndyS: any thoughts on how I should link "current version" for SPARQL docs when we don't yet have a pubdate?
16:54:44 [kasei]
or is that something I leave for the pub process?
16:55:07 [AndyS]
http://www.w3.org/TR/sparql11-query/ -- the TR undated is "latest"
16:55:49 [kasei]
right. but some of the refs use the "current [at time of publication] version" link.
16:55:57 [kasei]
(the dated-URI version)
16:56:50 [AndyS]
Not sure of the best thing to do - needs fix up in pub process.
16:57:10 [kasei]
maybe I'll just leave the link distinction off for now, and sort it out in pub process.
16:57:13 [kasei]
thanks
16:59:41 [ericP]
AndyS, i had to go to another call
17:00:37 [ericP]
AndyS, the goal is to make the editors' drafts as machine-checkable as possible?
17:01:01 [ericP]
yeah, leave that out
17:01:15 [ericP]
team contact will fix up the header
17:02:58 [ericP]
so e.g. /TR/2011/wd-sparql11-2011-12-25/ needs to reference /TR/2011/wd-sparql-update-2011-12-25/ without knowning "12-25" yet?
17:07:48 [kasei]
ericP: yes, that was the question.
17:10:07 [kasei]
AndyS: we've agreed to the changes in escaping and encoding in qnames, right? I assume there's need for new syntax tests?
17:11:01 [ericP]
for multi-doc suites, editors frequently count on xml-spec or some new JS thing which i forget the name of
17:11:44 [AndyS]
Ideally. Not by tomorrow though. I am uncomfortable with skipping CR for exactly this sort of issue - not doing what is needed because of mistakes. And anyway, graph store protocol now can't skip CR IMHO.
17:11:47 [ericP]
but if you pick a date, the team contact can grep for it and be pretty confident that he's gotton the links which need updating
17:11:55 [AndyS]
respec.js
17:12:01 [ericP]
yeah, that
17:12:22 [AndyS]
neither addresses greg's issue of coord.
17:12:59 [AndyS]
(well, not as far as I have found)
17:14:29 [chimezie]
chimezie has joined #sparql
17:27:37 [kasei]
AndyS: does ARQ dev support the new qname syntax?
17:27:49 [AndyS]
yes
17:27:56 [kasei]
I believe my impl does, but would like something else to test against if I try to crank out some tests.
17:28:04 [kasei]
great
17:31:16 [chimezie_]
chimezie_ has joined #sparql
18:01:48 [chimezie]
chimezie has joined #sparql
18:47:44 [chimezie]
chimezie has joined #sparql
18:57:23 [AxelPolleres]
AxelPolleres has joined #sparql
18:57:45 [AxelPolleres]
AxelPolleres has left #sparql
19:00:04 [chimezie_]
chimezie_ has joined #sparql
19:04:24 [chimezie]
chimezie has joined #sparql
19:39:27 [chimezie]
chimezie has joined #sparql
19:55:12 [alessandro]
alessandro has joined #sparql
19:56:28 [Zakim]
Zakim has left #sparql
19:57:24 [alessandro]
is it possible in regex for the pattern to be of the form "?a" ?
20:02:07 [chimezie_]
chimezie_ has joined #sparql
22:05:43 [AndyS]
alessandro - if you mean regex(?var, ?pattern) yes (it's just three args to regex -- normal args, inc ?v) - if you mean regex(?var, "Something|?pattern") then no -- no substituion into the pattern.
22:07:30 [alessandro_]
alessandro_ has joined #sparql
22:24:18 [AndyS]
AndyS has joined #sparql
22:35:27 [swh]
swh has joined #sparql
23:06:00 [pgearon]
pgearon has joined #sparql
23:39:33 [chimezie]
chimezie has joined #sparql