Chatlog 2009-12-01

From SPARQL Working Group
Jump to: navigation, search

See original RRSAgent log and preview nicely formatted version.

Please justify/explain all edits to this page, in your "edit summary" text.

14:57:00 <AndyS> Present: AxelPolleres, AndyS, kasei, SteveH, LukeWM, KjetilK, pgearon, dcharbon2, Chimezie_Ogbuji, LeeF, Ivan Herman, john-l, AlexPassant, Orri, Prateek
14:57:34 <RRSAgent> RRSAgent has joined #sparql
14:57:34 <RRSAgent> logging to
14:57:45 <Zakim> Zakim has joined #sparql
14:58:00 <AxelPolleres> Zakim, this will be SPARQL
14:58:01 <Zakim> "SPARQL" matches Team_(sparql-chairs)14:00Z, and SW_(SPARQL)10:00AM, AxelPolleres
14:59:38 <dcharbon2> dcharbon2 has joined #sparql
15:00:08 <chimezie> chimezie has joined #sparql
15:00:36 <AndyS> zakim, who is on the phone?
15:00:36 <Zakim> sorry, AndyS, I don't know what conference this is
15:00:37 <Zakim> On IRC I see chimezie, dcharbon2, Zakim, RRSAgent, AxelPolleres, LukeWM, AndyS, ivan, karl, pgearon, KjetilK, sandro, kasei, AlexPassant, trackbot, iv_an_ru
15:00:39 <SteveH_> SteveH_ has joined #sparql
15:00:52 <chimezie> Zakim, passcode
15:00:52 <Zakim> I don't understand 'passcode', chimezie
15:00:55 <AndyS> zakim, this is SPARQL
15:00:55 <Zakim> ok, AndyS; that matches SW_(SPARQL)10:00AM
15:01:03 <chimezie> Zakim, what is the passcode?
15:01:03 <Zakim> the conference code is 77277 (tel:+1.617.761.6200 tel:+ tel:+44.117.370.6152), chimezie
15:01:08 <LukeWM> zakim, Garlik has LukeWM,SteveH
15:01:10 <Zakim> sorry, LukeWM, I do not recognize a party named 'Garlik'
15:01:18 <AndyS> zakim, who is on the phone?
15:01:18 <Zakim> On the phone I see +03539149aaaa, ??P6, +1.310.729.aabb, ??P8, +1.312.863.aacc
15:01:20 <Zakim> + +1.919.543.aadd
15:01:24 <AxelPolleres> Zakim, who is on the phone?
15:01:24 <Zakim> On the phone I see +03539149aaaa, ??P6, +1.310.729.aabb, ??P8, +1.312.863.aacc, +1.919.543.aadd
15:01:26 <kasei> Zakim, aabb is me
15:01:26 <Zakim> +kasei; got it
15:01:27 <AndyS> zakim, ??P6 is me
15:01:27 <Zakim> +AndyS; got it
15:01:38 <AxelPolleres> Zakim, aaaa is me
15:01:38 <Zakim> +AxelPolleres; got it
15:01:39 <Zakim> +Chimezie_Ogbuji
15:01:44 <pgearon> zakim, aacc is me
15:01:44 <Zakim> +pgearon; got it
15:01:45 <dcharbon2> zakim, +1.919.543 is me
15:01:45 <Zakim> +dcharbon2; got it
15:01:52 <AxelPolleres> Zakim, who is on the phone?
15:01:52 <Zakim> On the phone I see AxelPolleres, AndyS, kasei, ??P8, pgearon, dcharbon2, Chimezie_Ogbuji
15:01:55 <SteveH_> what's the code to put your hand up?
15:01:59 <Zakim> +??P13
15:02:05 <AlexPassant> Zakim, ??p13 is me
15:02:05 <Zakim> +AlexPassant; got it
15:02:13 <kasei> 41#, I think
15:02:28 <SteveH_> Zakim, ??P8 is [Garlik
15:02:28 <Zakim> +[Garlik; got it
15:02:32 <SteveH_> Zakim, ??P8 is [Garlik]
15:02:32 <Zakim> I already had ??P8 as [Garlik, SteveH_
15:02:41 <Zakim> +[IPcaller]
15:02:48 <SteveH_> Zakim, [Garlik has SteveH_ and LukeWM 
15:02:48 <Zakim> +SteveH_, LukeWM; got it
15:02:49 <ivan> zakim, dial ivan-voip
15:02:49 <Zakim> ok, ivan; the call is being made
15:02:50 <Zakim> +Ivan
15:02:51 <AxelPolleres> AxelPolleres has changed the topic to: Agen�da -
15:03:01 <AndyS> zakim, [IPCaller] is Orri
15:03:01 <Zakim> +Orri; got it
15:03:18 <AndyS> Scribe: AndyS
15:03:49 <AndyS> Missing -- actions tracking:
15:04:22 <Zakim> +??P17
15:04:24 <john-l> john-l has joined #sparql
15:04:31 <KjetilK> Zakim, ??P17 is me
15:04:31 <Zakim> +KjetilK; got it
15:04:39 <KjetilK> Zakim, mute me
15:04:39 <Zakim> KjetilK should now be muted
15:05:03 <AxelPolleres> ACTION: axel to put minutes approval on agenda for next time... action need to be added/checked
15:05:03 <trackbot> Created ACTION-139 - Put minutes approval on agenda for next time... action need to be added/checked [on Axel Polleres - due 2009-12-08].
15:05:27 <AndyS> Next meeting - same time next week.
15:05:45 <AndyS> Axel: Scribe wil be Luke.
15:05:46 <Zakim> + +1.216.445.aaee
15:05:52 <john-l> Zakim: aaee is me
15:05:57 <AxelPolleres> scribe next time: lee (luke on the deck)
15:06:00 <john-l> Zakim, aaee is me
15:06:00 <Zakim> +john-l; got it
15:06:16 <AndyS> Topic: Comments handling
15:06:47 <AndyS> Axel: Unclear as to process decided last time.
15:07:14 <AndyS> ... ?? Greg to follow up on ldodds / service description comment
15:07:29 <AndyS> kasei: sure
15:07:38 <AxelPolleres> ACTION: axel to track all mails on comments list on wiki.
15:07:38 <trackbot> Created ACTION-140 - Track all mails on comments list on wiki. [on Axel Polleres - due 2009-12-08].
15:07:55 <AxelPolleres>
15:08:05 <AndyS> SubTopic: Comment:BL
15:08:27 <AndyS> Axel: any comments on the proposed response?
15:08:40 <AndyS> ... any objections to text on wiki?
15:08:55 <AxelPolleres> ACTION: axel to send out response to BL
15:08:55 <trackbot> Created ACTION-141 - Send out response to BL [on Axel Polleres - due 2009-12-08].
15:09:43 <AndyS> SubTopic: Toby Inkster comment - email dialogue started
15:10:17 <AndyS> Axel: May be better to follow process: first wiki draft a reposne, WG approves, then send
15:10:23 <SteveH_> sounds ok
15:10:27 <AlexPassant> ok
15:10:58 <AndyS> Topic: Comments process
15:11:10 <AndyS> Axel: seek acknowledgement of response
15:11:19 <Prateek> Prateek has joined #sparql
15:11:25 <AxelPolleres> status - open/RESPONED/RESP+ACK
15:11:29 <AlexPassant> can we use the tracker for that ?
15:11:48 <Zakim> + +1.937.775.aaff
15:11:59 <Prateek> Zakim,+1.937.775.aaff is Prateek
15:11:59 <Zakim> +Prateek; got it
15:12:11 <AndyS> q+ to ask how things get onto the wiki in the first place
15:12:53 <AndyS> ivan: tracker makes work later
15:13:07 <AndyS> ... wiki is ready for the review processing
15:13:33 <AxelPolleres> ack AndyS
15:13:33 <Zakim> AndyS, you wanted to ask how things get onto the wiki in the first place
15:13:37 <SteveH_> ack ??P8
15:13:50 <AxelPolleres> Zakim, unmute AndyS
15:13:50 <Zakim> AndyS was not muted, AxelPolleres
15:14:02 <AndyS> seem to be muted by technology
15:14:20 <AndyS> How do comments get onto the wiki?  axel to trck comments email list?
15:14:46 <AndyS> axel: yes
15:15:21 <AndyS> Topic: Liasons
15:15:46 <AndyS> Orri: RDF2RDB - reviewing existing work
15:16:00 <AndyS> ... support vendor specific extensions
15:16:12 <AndyS> ... more to come - couple of weeks
15:16:36 <AndyS> Steve: eGov - nothing to report
15:16:43 <AndyS> Regrets: sandro, BirteG
15:18:29 <AndyS> Axel: What were the conclusions about Service Descriptions?
15:18:44 <AndyS> .. taken to email?  Didn't see email.
15:19:12 <AndyS> kasei: changes made in CVS
15:19:25 <KjetilK> q+ to say where DELETE abbrev stands
15:19:32 <AxelPolleres> Action: greg to summarise changes of SD and conclusions from last TC along with open questions.
15:19:32 <trackbot> Created ACTION-142 - Summarise changes of SD and conclusions from last TC along with open questions. [on Gregory Williams - due 2009-12-08].
15:19:39 <KjetilK> ack me
15:19:40 <Zakim> KjetilK, you wanted to say where DELETE abbrev stands
15:20:03 <AndyS> Topic: Abbrev DELETE
15:20:28 <AndyS> kjetilk: need to consider wider context
15:20:41 <AndyS> ... but WG keen to have abbrev DELETE
15:20:43 <KjetilK> Zakim, mute me
15:20:43 <Zakim> KjetilK should now be muted
15:20:52 <AndyS> Axel: will cover today
15:21:10 <AndyS> ... and relates to dataset/graph store
15:21:14 <AxelPolleres>
15:21:21 <AndyS> Topic:  dataset/graph store
15:21:31 <AndyS> Axel: Will summarise
15:21:45 <AxelPolleres>
15:21:59 <AndyS> ... At F2F1, issue-51, meaning of dataset description clauses - allowed?
15:22:37 <AndyS> ... Graph store = collection of graphs but != dataset
15:22:53 <AndyS> ... e.g. graph URI used that is not in graph store 
15:23:03 <LeeF> LeeF has joined #sparql
15:23:52 <AndyS> ... Lee summary: pattern mathicn against graph store - no restrictions
15:24:02 <AndyS> q+
15:25:00 <AndyS> AndyS: don't see a use case for dataset description
15:25:43 <AndyS> ... DELETE FROM overloads FROM with a different meaning
15:25:52 <AxelPolleres> DELETE FROM <http://example/bookStore>
15:25:53 <AxelPolleres>  { ?book ?p ?v }
15:25:53 <AxelPolleres> WHERE
15:25:53 <AxelPolleres>   {
15:26:03 <AxelPolleres> DELETE {GRAPH <http://example/bookStore>
15:26:03 <AxelPolleres>  { ?book ?p ?v }}
15:26:03 <AxelPolleres> WHERE
15:26:03 <AxelPolleres>   {
15:26:22 <SteveH_> -1 to DELETE FROM
15:26:31 <SteveH_> +1 to DELETE ... GRAPH
15:26:36 <AndyS> (PaulG wrote about this last night)
15:27:33 <AndyS> pgearon: don't see a difference - syntactic variant
15:27:43 <SteveH_> INSERT { GRAPH <x> { ... } GRAPH <y> { ... } }
15:28:32 <AndyS> pgearon: quads blocks don't look like SQL which is my (paul) pref 
15:28:33 <AxelPolleres> INSERT INTO and DELETE FROM as syntactic sugar? for INSERT/DELETE { GRAPH ...
15:28:36 <KjetilK> SteveH_, so, moving triples between graphs?
15:28:42 <dcharbon2> q+
15:28:45 <LeeF> I'm happy to hvae GRAPH { ... } specify the target of what gets inserted/deleted, but i'm not happy to lose the ability to use FROM & FROM NAMED to define an RDF dataset for the WHERE part of an update operation
15:29:00 <SteveH_> I don't like the idea of overloading FROM
15:29:03 <KjetilK> q+
15:29:18 <AndyS> LeeF, it's not lost - it's not there currently as far as I can see.
15:29:27 <LeeF> AndyS, exactly
15:29:38 <LeeF> lose it when going from query -> update
15:29:48 <AndyS> Leef, is it needed?  Use case?
15:30:01 <AndyS> pgearon: no WITH means apply to unnamed graph
15:30:05 <LeeF> anytime I want to scope the graphs that my insert/delete pattern-matching operate over
15:30:15 <AndyS> Axel: every graph store has a default graph store?
15:30:41 <AndyS> pgearon: if ref to non-existant graph then error just like <NotInGraphStore>
15:31:01 <LeeF> (basically, the same use cases as FROM & FROM NAMED in SPARQL Query)
15:31:02 <AndyS> axel: update in service description?
15:31:21 <AndyS> kasei: not in SD
15:31:24 <SteveH_> IIRC you're required to have an unnamed graph, though it can be empty of course [splitting hairs
15:31:52 <AxelPolleres> do we need the availability of  an unnamed graph in SD? (issue)
15:31:52 <trackbot> Created ISSUE-52 - Do we need the availability of  an unnamed graph in SD? ; please complete additional details at .
15:32:13 <AndyS> LeeF, OK - access to graphs not in the store on web etc
15:32:54 <AndyS> pgearon: NO FROM,INTO or WITH is default graph
15:33:00 <AxelPolleres> I am actually not 100% sure whether we haven't got already an issue about that (whether or not an unnnamed graph is required/implicit?)
15:33:14 <LeeF> AndyS, well, sure, but in my case it's really more about constraining the scope of the pattern matching to a union of a subset of the graphs in my store
15:33:29 <SteveH_> "a Graph Store contains one unnamed graph and zero or more named graphs"
15:33:30 <AndyS> LeeF, what about GRAPH like query uses?
15:33:34 <SteveH_>
15:33:51 <AndyS> LeeF, take to email?
15:34:26 <LeeF> AndyS, sure
15:34:34 <AxelPolleres> ACTION: axel link issue-52 to issue-20 and check back
15:34:34 <trackbot> Created ACTION-143 - Link issue-52 to issue-20 and check back [on Axel Polleres - due 2009-12-08].
15:34:40 <KjetilK> q?
15:34:50 <AndyS> q-
15:35:31 <AndyS> There is an unnamed graph always (it can be empty)- make it like RDF dataset.
15:36:07 <SteveH_> q+
15:36:16 <KjetilK> ack dcharbon
15:36:38 <AndyS> dcharbon: observe DELETE FROM like SQL but could specify before in request (slice of graph store foro rest of request until next slice)
15:37:03 <AndyS> ... more like  choose subset ... actions .. choose subset ... 
15:37:04 <chimezie> as in:
15:37:10 <chimezie> FROM ....
15:37:17 <chimezie> DELETE { .... }
15:37:40 <SteveH_> q?
15:37:48 <KjetilK> ack me
15:38:48 <AxelPolleres> seems one main crux is issue-20, the other is whether we want syntactic sugar FROM/INTO/(USING)
15:38:49 <SteveH_> ack me
15:38:58 <AndyS> kjetilk: arg against DELETE FROM is that FROM is uses elsewhere (query)
15:39:19 <AndyS> SteveH_: false impression for SQL ; confusing for SPARQL
15:39:42 <AndyS> q+
15:39:56 <SteveH_> +1 to AndyS 
15:40:09 <SteveH_> AndyS: I don't know what syntactic sugar
15:40:37 <KjetilK> Zakim, mute me
15:40:37 <Zakim> KjetilK should now be muted
15:40:46 <AndyS> ack me
15:40:47 <chimezie> using GRAPH with INSERT seems odd
15:41:22 <KjetilK> +1 (should either do only DELETE FROM or GRAPH, not syntactic sugar as the former for the latter)
15:41:41 <AxelPolleres> strawpoll ... shall we drop DELETE FROM/ INSERT INTO, in favor of just having DELETE/INSERT { GRAPH ...
15:42:23 <AndyS> chimezie: all or nothing question.  Find INSERT INTO natural
15:42:23 <AxelPolleres> shall we drop DELETE FROM ?
15:42:29 <SteveH_> +1
15:42:32 <chimezie> +1
15:42:39 <AxelPolleres> 0
15:42:42 <AlexPassant> +1
15:42:43 <ivan> 0
15:42:45 <LukeWM> 0
15:42:47 <dcharbon2> 0
15:42:50 <pgearon> -1
15:42:56 <AndyS> Can't say
15:43:02 <kasei> +1
15:43:16 <AndyS> My pref (0.75) is GRAPH uniformly. 
15:43:37 <AndyS> Either INTO and FROM or GRAPH - mixed is weird.
15:43:43 <SteveH_> [verb] FROM ... WHERE { }
15:43:57 <SteveH_> +1 to AndyS 
15:44:46 <AxelPolleres> shall we drop INSERT INTO?
15:44:48 <LukeWM> +1
15:44:49 <SteveH_> +1
15:44:56 <ivan> 0
15:45:02 <kasei> +1
15:45:02 <pgearon> only if dropping DELETE FROM
15:45:05 <AxelPolleres> 0
15:45:11 <KjetilK> +1 to pgearon
15:45:13 <chimezie> if the alternative is INSERT {GRAPH .. {..}} that would seem very confusing to me, so -1
15:45:15 <AlexPassant> +1 to pgearon 
15:45:19 <Zakim> -dcharbon2
15:45:24 <AndyS> +1 to track with DELETE FROM
15:46:07 <AndyS> Is there a non-FROM word we can use?
15:46:22 <AndyS> chimezie: use cases would be good
15:46:36 <AndyS> +1 to chimezie
15:46:40 <SteveH_> INSERT { GRAPH ?x { ?x ?y ?z } } WHERE { ?x ?y ?z } is a usecase
15:46:47 <pgearon> +1 to use cases
15:46:57 <AndyS> INSERT { GRAPH ?g ...  :-)
15:47:16 <AndyS> pgearon: feel there is not consensus on direction
15:47:18 <Zakim> +dcharbon2
15:47:28 <AndyS> axel: who will write uses cases?
15:47:49 <AndyS> pgearon: I offer to start discussion
15:47:53 <AxelPolleres> ACTION: paul to start conversation in an email
15:47:53 <trackbot> Created ACTION-144 - Start conversation in an email [on Paul Gearon - due 2009-12-08].
15:48:04 <KjetilK> q+
15:48:18 <AndyS> ACTION-144 - paul to start conversation in an email about uses cases for INSERT / DELETE
15:48:18 <trackbot> ACTION-144 Start conversation in an email notes added
15:48:35 <KjetilK> ack me
15:48:40 <KjetilK>
15:48:57 <AndyS> kjetilk: remind people we have use cases (see link)
15:48:57 <KjetilK> Zakim, mute me
15:48:57 <Zakim> KjetilK should now be muted
15:49:37 <AndyS> Those use caes don't compare and contrast the use of the two styles.
15:49:51 <AndyS> pgearon: want to have cases with multiple graphs
15:50:02 <AndyS> ... start with wiki examples
15:50:07 <KjetilK> We have one that may be relevant:
15:50:10 <LukeWM> q+
15:50:19 <AndyS> axel: INSERT INTO one graph at a time
15:50:36 <AndyS> pgearon: quads makes it hard for triple stores
15:50:36 <LukeWM> ack me
15:50:59 <chimezie> In that use case (moving between graphs) the use of GRAPH/INTO seem more intuitve
15:50:59 <AndyS> ... want to have the ability to match onegraph and use template to insert into another
15:51:05 <chimezie> s/intuitve/intuitive
15:51:23 <LukeWM> LeeF's email had an example of using FROM in insert:
15:51:29 <AndyS> (also applies to DELETE from a graph by template)
15:51:51 <AndyS> axel: Variables in FROM?
15:52:02 <AndyS> chimezie: what's the var scope?
15:52:13 <AndyS> .. and range
15:52:32 <AndyS> ... or does the WHERE apply.
15:52:47 <AndyS> axel: is there a use case for insert into every graph in store?
15:52:58 <AndyS> pgearon: what about insert ontology everywhere?
15:53:05 <AndyS> orri: loop on variable
15:53:07 <SteveH_> INSERT { GRAPH ?g { <a> <b> <c> } WHERE { GRAPH ?g {} }
15:53:14 <SteveH_> don't need unbound variables
15:53:55 <AndyS> axel: what if var is boud to URI not namign a graph store?  error? skip? create?
15:53:57 <AxelPolleres> insert { graph ?g { } }
15:54:06 <AndyS> q+
15:54:42 <AxelPolleres> means inserting into the graphs with bindings for ?g, if the graph doesn't exist, then create it...
15:55:02 <AxelPolleres> andyS: this is problematic, if the graphstore only alllows certain URIs
15:56:02 <AndyS> orri: we would create.  But otehr may generate error.
15:56:11 <SteveH_> that would be analagous to a URI that you just didn't like e.g. <a\tb>
15:56:14 <SteveH_> surely?
15:56:43 <dcharbon2> how do I delete from 3 named graphs without from? like this? delete { graph ?g { <s> <p> <o> } } where { graph ?g {} filter (?g = <g1> || ?g = <g2> || ?g = <g3>) }
15:56:44 <AxelPolleres> ACTION: axel to check whether we have an issue for insert into ?var, otherwise create it.
15:56:44 <trackbot> Created ACTION-145 - Check whether we have an issue for insert into ?var, otherwise create it. [on Axel Polleres - due 2009-12-08].
15:56:50 <AndyS> SteveH_, that might be (shoudl be?) trapped earlier.
15:57:05 <SteveH_> AndyS, perhaps
15:57:06 <pgearon> I'd really like the option of making an unknown graph an error. I can see a case where a bug could pollute a graph store with thousands of unwanted graphs
15:57:11 <AndyS> "it" = issue not the graph
15:57:46 <AndyS>  { graph <a> { <s> <p> <o> } graph <b> { <s> <p> <o> } graph <c> { <s> <p> <o> } }
15:58:33 <AndyS> axel: take to email
15:59:18 <AndyS> pgearon: lack of DROP equivalent is worrying.
15:59:35 <AndyS> ... can't do a mass removal of graphs.
15:59:52 <SteveH_> INSERT { ?x ?y ?z  WHERE { GRAPH ?g { ?x ?y ?z } } is just as destructive,and can't be undone
16:00:12 <AndyS> This is an issue - is a graph with no triples exist
16:00:39 <AxelPolleres> existence of empty graph is again issue-20, BTW
16:00:59 <AndyS> Issue owner?
16:01:09 <AndyS> pgearon: different for triple and quad stores
16:01:46 <AndyS> ... INSERT creates, DELETE does not drop
16:01:58 <SteveH_> DROP ?g WHERE { GRAPH ?g {} }
16:01:59 <SteveH_> ?
16:02:08 <AndyS> We must support both styles - e.g. graph metadata
16:02:18 <dcharbon2> +1 SteveH_
16:02:21 <AndyS> q+
16:02:26 <AxelPolleres> implicit CREATE is assymetric to DELETE which doesn't necessarily drop.
16:03:13 <AxelPolleres> AndyS: another use case is what graphs are in my graph store.
16:03:29 <AndyS> Axel: core is issue-20
16:03:43 <AndyS> ... who would lead this?
16:04:01 <pgearon> I'm also uncomfortable with having to monitor all inserts to see if I need to identify new graphs being created (which I need to do for our graph management)
16:04:37 <AndyS> axel: overtime now - will summarise on email
16:04:55 <Zakim> -Chimezie_Ogbuji
16:04:56 <Zakim> -Prateek
16:04:57 <Zakim> -Orri
16:04:59 <Zakim> -john-l
16:05:01 <Zakim> -Ivan
16:05:01 <AndyS> We didn't get to aggregate functions
16:05:02 <Zakim> -AlexPassant
16:05:03 <Zakim> -kasei
16:05:03 <Zakim> -[Garlik
16:05:05 <Zakim> -dcharbon2
16:05:08 <AndyS> ADJOURNED
16:05:24 <AxelPolleres> ACTION: Axel to make a draft on  alternatives for ISSUE-20
16:05:24 <trackbot> Created ACTION-146 - Make a draft on  alternatives for ISSUE-20 [on Axel Polleres - due 2009-12-08].
16:05:38 <Zakim> -pgearon
16:06:47 <Zakim> -AndyS
16:06:48 <AxelPolleres> rrsagent, make records public
16:07:09 <AxelPolleres> andy, will you take care of getting the minutes online?
16:08:36 <Zakim> -KjetilK
16:09:30 <AndyS> Wil try to do the minutes
16:10:03 <AxelPolleres> so, were there really no actions last time, or were they just not recorded... if there were, it seems they go lost in the IRC log... :-(
16:10:06 <AxelPolleres> ?
16:10:12 <AxelPolleres> thanks, andy!