15:00:46 <RRSAgent> RRSAgent has joined #sparql
15:00:46 <RRSAgent> logging to http://www.w3.org/2010/07/30-sparql-irc
15:00:52 <LeeF> zakim, room for 10 for 90 minutes?
15:00:54 <Zakim> ok, LeeF; conference Team_(sparql)15:00Z scheduled with code 772775 (SPARQL) for 90 minutes until 1630Z
15:00:57 <LeeF> thanks, zakim
15:01:08 <LeeF> RRSAgent, make logs world
15:01:18 <LeeF> Meeting: SPARQL WG dedicated meeting on formal update model
15:01:20 <bglimm> Zakim, passcode?
15:01:20 <Zakim> the conference code is 772775 (tel:+1.617.761.6200 tel:+33.4.26.46.79.03 tel:+44.203.318.0479), bglimm
15:01:22 <Zakim> Team_(sparql)15:00Z has now started
15:01:29 <Zakim> +[IPcaller]
15:01:34 <AndyS> zkim, [IPCaller] is me
15:01:35 <Zakim> +pgearon
15:01:38 <AndyS> zakim, [IPCaller] is me
15:01:38 <Zakim> +AndyS; got it
15:01:46 <Zakim> +??P13
15:01:47 <Zakim> +Lee_Feigenbaum
15:02:00 <Zakim> +bglimm
15:02:06 <AlexPassant> Zakim, ??P13 is me
15:02:06 <Zakim> +AlexPassant; got it
15:02:11 <bglimm> Zakim, mute me
15:02:11 <Zakim> bglimm should now be muted
15:02:23 <Zakim> +Sandro
15:02:27 <Zakim> +[IPcaller]
15:02:39 <kasei> Zakim, [IPcaller] is me
15:02:39 <Zakim> +kasei; got it
15:03:02 <LeeF> zakim, who's here?
15:03:02 <Zakim> On the phone I see AndyS, pgearon, Lee_Feigenbaum, AlexPassant, bglimm (muted), Sandro, kasei
15:03:04 <Zakim> On IRC I see RRSAgent, Zakim, AndyS, bglimm, LeeF, SteveH, AlexPassant, pgearon, iv_an_ru, karl, sandro, trackbot, kasei
15:03:07 <AxelPolleres> AxelPolleres has joined #sparql
15:03:26 <LeeF> Scribenick: AndyS
15:03:56 <AndyS> LeeF: it's Friday ...
15:04:33 <AndyS> ... intent: find the approach for a formal model of update
15:04:50 <Zakim> +AxelPolleres
15:04:54 <AxelPolleres> Zakim, who is on the phone?
15:04:54 <Zakim> On the phone I see AndyS, pgearon, Lee_Feigenbaum, AlexPassant, bglimm (muted), Sandro, kasei, AxelPolleres
15:05:05 <AndyS> ... goal - leave meeting with material for Paul and Alex (or other volunteer) to craft text
15:05:37 <LeeF> Lee: how does query's model work?
15:05:44 <LeeF> AndyS: you have a string in your hand - what is the right answer to come out?
15:05:54 <LeeF> ... 1. parse the syntax into an abstract syntax tree
15:06:12 <LeeF> ... (expands out syntax shortcuts, etc.)
15:06:28 <LeeF> ... 2. run a fixed algorithm to turn the abstract syntax into the algebra
15:06:32 <LeeF> ... 3. evaluate the algebra
15:06:44 <Zakim> +[IPcaller]
15:06:46 <LeeF> ... somewhat repetitive - close relationhsip between the terms in the algebra and their evaluation
15:06:51 <SteveH> Zakim, [IPcaller] is me
15:06:51 <Zakim> +SteveH; got it
15:07:50 <AndyS> LeeF - approaches include model theory, procedural style in SPARQL, other
15:08:31 <AndyS> ... suggestion: take the procedural style for SPARQL Update.
15:08:55 <AxelPolleres> in principle, sounds fine, depends on how we define "procedural"
15:09:20 <AxelPolleres> +1 to update mapping from one set/state to another 
15:09:23 <AndyS> Paul, update are a mapping from GS to GS, variations of this (GS-1,2,3)
15:09:32 <AndyS> Paul: update are a mapping from GS to GS, variations of this (GS-1,2,3)
15:09:45 <AndyS> ... implementation follow the data structure style
15:10:07 <AndyS> .. there is potential disconnect there
15:10:19 <AndyS> LeeF: experience, prefer functional style
15:11:01 <AndyS> Input - http://www.w3.org/2009/sparql/wiki/Lees_Update_Graph_Model
15:11:15 <AxelPolleres> have put some draft of what I think, which apart from the redefinition of Dataset should work in my mail http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JulSep/0126.html I think Andy had some similar proposals in another mail... can't find it at the moment.
15:11:33 <AndyS> Input - http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JulSep/0127.html
15:12:07 <AndyS> Paul: Mulgara is immutable structures for transactionality (single writer)
15:12:22 <AxelPolleres> Andy, I think you had something more concrete on the semantics of various operations in some earlier mail, can't find it at the moment, maybe I misremember
15:12:40 <AndyS> LeeF: Spec helps for test cases - links formal defns to practical implementation
15:12:43 <AxelPolleres> s/various operations/various update operations/
15:13:14 <AndyS> /em coudl yo ufind it?  was it  http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JulSep/0125.html  [scribing...]
15:13:48 <AxelPolleres> q+
15:14:05 <AndyS> LeeF: is an update operation from GS to GS?
15:14:51 <AxelPolleres> I thought  "Graphstore" could also use the definition of dataset, meaning that a graphstore
15:14:51 <AxelPolleres> is defined by a sequence of datasets DS_0 to DS_n determined by a sequence of 
15:14:52 <AxelPolleres> updates: Starting with DS_0 being the empty or initial dataset, 
15:14:52 <AxelPolleres> the "current dataset" after the n-th update is simply DS_n.
15:15:04 <AndyS> Sandro: what about GS-state to GS-state?
15:15:10 <LeeF> service <-> graphstore <-> state1 -> state2 -> state3 etc.
15:15:19 <LeeF> ack AxelPolleres
15:17:22 <AndyS> Axel: key is define state DS(n-1) -> DS(n)
15:17:43 <AndyS> LeeF: what is the state of a GS?  is it a DS?
15:18:08 <AxelPolleres> Axel: thinks that the state of a graph store should be representable as an RDF dataset
15:18:46 <LeeF> Is a GS-state the same as an RDF dataset? a pair of (graph, Set of Pairs of (Name, Graph))
15:18:50 <AlexPassant> "state of a GS is a DS at a particular time" might be better
15:19:04 <AxelPolleres> alex, works for me.
15:19:05 <AlexPassant> (graph, set of pairs, T)
15:19:23 <AlexPassant> so we have (graph, set of pairs, Ta) -> operation -> (graph, set of paris, Tb)
15:20:05 <AxelPolleres> I used T_n T_n+1 to indicate that I mean that to be atomic, i.e time here being meant discrete, nothing happening in between
15:20:21 <AndyS> LeeF: do we agree that the state of a GS is a DS or similar?
15:20:25 <pgearon> I'm happy with that definition
15:20:53 <AlexPassant> agreed with that AxelPolleres (re. atomicity)
15:21:31 <sandro> +1 this meeting is gathering advice for editors, not constraints for them.
15:21:45 <AndyS> LeeF: What is the concept of the time here?
15:22:16 <AndyS> Alex: discrete time, point-wise transitions to new state
15:22:58 <LeeF> GS - graph store
15:23:02 <LeeF> GSS - graph store state
15:23:05 <LeeF> DS - rdf dataset
15:23:12 <LeeF> Op - Update operation
15:23:15 <LeeF> Req - Update request
15:23:18 <AxelPolleres> rather, a graphstore is a "state machine" that moves from one state to another by update operations.
15:23:30 <LeeF> Op - a function from GSS -> GSS
15:23:48 <AndyS> That is the service that is the state machine?  Or is that the same?
15:24:01 <AxelPolleres> so, op is the state transition function, which semantics we have to define in terms of pre and post state
15:24:06 <LeeF> Req is Op1, Op2, Op3 -  GSS -> GSS - Op3(Op2(Op1(GSS0)))
15:24:23 <AndyS> +1
15:24:31 <pgearon> +1
15:24:35 <sandro> +1
15:24:36 <AxelPolleres> looks ok
15:24:37 <AlexPassant> +1
15:24:59 <AndyS> (it says to me the Req is atomic ++)
15:25:29 <LeeF> LOAD - http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml#t417
15:25:43 <LeeF> "http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml#t417"
15:25:52 <LeeF> LOAD <documentURI> [ INTO GRAPH <uri> ]
15:26:14 <AxelPolleres> LOAD <documentURI> [ INTO GRAPH <gnew> ]
15:26:14 <AxelPolleres>  this means
15:26:14 <AxelPolleres>  DS_{t+1} = DS_t union {(<gnew>_t, G_<gnew>)} 
15:26:26 <AxelPolleres> something like that...
15:27:36 <SteveH> why <gnew>_t ?
15:27:38 <AndyS> DS1 is DS[0 with G replaced by GET <uri> 
15:27:57 <AxelPolleres> DS_{t+1} = DS_t union {(<gnew>, G_<gnew>)}
15:28:05 <AndyS> .. notation for graph in datatset?  DS[g]?
15:28:31 <AxelPolleres> DS_{t+1} = DS_t union {(<gnew>, G_<documentURI>)}
15:28:41 <AndyS> DS_{t+1} = DS_t \ {(<v>, old)} union {(<gnew>, G_<gnew>)}
15:28:47 <AxelPolleres> if graph is already there...
15:28:58 <AxelPolleres> LOAD <documentURI> [ INTO GRAPH <g> ]
15:29:00 <AxelPolleres> means 
15:29:23 <AndyS> just take the old one out regard less with \ {(<g>, any)}
15:29:34 <SteveH> +1
15:29:48 <AxelPolleres> DS_t { .... (<g>, G) ... }
15:29:50 <AxelPolleres> =>
15:30:17 <SteveH> some people have triplestore, not quadstores
15:30:31 <AxelPolleres> DS_t+1  =  {    (<g>, G merge G_<documentURI>)   }
15:30:37 <LeeF> GRAPH ?g { } # test if graph exists
15:30:40 <AxelPolleres> or alike
15:31:03 <LeeF> Sandro: is the difference between pairs and quads the existence of empty graphs?
15:31:08 <LeeF> AndyS: yes
15:31:11 <LeeF> sandro: is this observable?
15:31:14 <LeeF> AndyS: yes
15:31:30 <Zakim> -AlexPassant
15:32:21 <AndyS> CREATE GRAPH <uri> should cause GRAPH ?g {} to show <uri>
15:32:35 <SteveH> AndyS, could, not should
15:33:25 <LeeF> sandro: a no-op for CREATE will pass whatever test we have since implementations are allowed to prune empty graphs, so we should be able to define the semantics using quads
15:33:44 <SteveH> I agree with sandro, about defining in terms of quads would work, but I'm not sure it's easier
15:34:28 <Zakim> +??P2
15:34:31 <AlexPassant> Zakim, ??P2 is me
15:34:31 <Zakim> +AlexPassant; got it
15:34:46 <AxelPolleres> do we need to devlare the behaviour of operations for both graph-aware and non-graph-aware stores?
15:35:35 <AndyS> Quads are an implmentation technique we should respect but the defn should be graphs as RDF is about graphs. 
15:35:56 <AndyS> Quads have problems with the default graph.
15:36:41 <LeeF> sandro: do we have any way in the test suite setup of saying that either of 2 results would be ok?
15:36:51 <LeeF> AndyS: we have 1 case of that right now in query
15:36:53 <AxelPolleres> q+ to ask/talk about update test cases
15:37:04 <LeeF> AndyS: there are 2 different tests to capture both possibilities
15:37:39 <LeeF> AndyS: we could divide places where they are different into 2 sets and implementers choose which set to run
15:37:44 <SteveH> +1 to not writing tests where their different :)
15:37:46 <LeeF> ... other option is not to write the tests where they are different
15:38:11 <LeeF> ack AxelPolleres
15:38:11 <Zakim> AxelPolleres, you wanted to ask/talk about update test cases
15:38:55 <AndyS> Another matter for quads is the lack of std quads format.
15:39:41 <AndyS> LeeF: could have two tests, one with pruning, one without
15:40:13 <AndyS> Axel: tests defined in state-before, ans state-after and mark tests "foro graph aware GS"
15:40:20 <AndyS> s/ans/and/
15:40:31 <AndyS> s/foro/for/
15:41:03 <AxelPolleres> SELECT ?S ?P ?O WHERE { ?S ?P ?O OPTIONAL { GRAPH ?G {?S ?P ?O } } } would be an (ugly) way to query the graphstore state, if we assume  graphstore tied to default dataset
15:41:15 <SteveH> ...what about if we said GRAPH ?g { } always returned nothing?
15:41:43 <iv_an_ru> +1
15:41:56 <LeeF> SteveH, that would be a nonstarter for me, I believe
15:42:00 <SteveH> ok
15:42:42 <LeeF> CREATE GRAPH G -->  if G exists, error or nothing ; else if GSS0 = (UG0, NGS0, T0) then CREATE(GSS0, G) = (UG0, NGS0 union (G, {}), T0+1)
15:42:50 <AxelPolleres> in test cases we might need to distunguish the behavior of graph-aware and non-graph-aware update-endpoints
15:43:10 <AndyS> Confusing if GRAPH ?g { } = nothing but GRAPH ?G {?S ?P ?O } something 
15:43:27 <SteveH> not really, but doesn't seem to be a popular idea
15:43:50 <iv_an_ru> AxelPolleres, agree re. tests.
15:44:21 <LeeF> AndyS: the way to do this is to define the operations with empty graphs, and then note that implicit pruning may go on
15:44:39 <SteveH> I don't like the term "graph aware" FWIW, I think it's confusing
15:44:56 <AndyS> Axel: section on "graph aware" and "graph pruning" stores
15:45:07 <SteveH> or two sets of results, either of which is ok
15:45:14 <AndyS> (better to mention graph graphs)
15:46:07 <AndyS> sandro: concern of interoperability in the market
15:47:29 <AndyS> LeeF: looks like there is some work to be done but general direction is appearing
15:48:10 <sandro> sandro: let's not formalize two classes of sparql end points (graph aware and non-graph aware), eg in the test suite, since doing so would fragment the market.
15:48:25 <AndyS> pattern -> solution sequence -> instantiate a sets of triples for each GRAPH mentioned (and dft graph) -> make changes
15:49:31 <pgearon> yes
15:49:32 <AlexPassant> sounds ok for me
15:50:26 <pgearon> missing considerations is my only concern at this point, but I can't think of any at the moment
15:50:32 <Zakim> -Sandro
15:50:34 <Zakim> -Lee_Feigenbaum
15:50:35 <Zakim> -bglimm
15:50:36 <Zakim> -AxelPolleres
15:50:36 <LeeF> AndyS, thanks for scribing
15:50:36 <pgearon> OK, thank you
15:50:36 <Zakim> -AlexPassant
15:50:49 <Zakim> -pgearon
15:50:53 <Zakim> -AndyS
15:50:54 <SteveH> bye
15:51:01 <Zakim> -SteveH
15:51:55 <Zakim> -kasei
15:51:56 <Zakim> Team_(sparql)15:00Z has ended
15:51:58 <Zakim> Attendees were pgearon, AndyS, Lee_Feigenbaum, bglimm, AlexPassant, Sandro, kasei, AxelPolleres, SteveH
16:05:43 <AxelPolleres> AxelPolleres has left #sparql
16:22:28 <AlexPassant> RRSAgent: create minutes
16:22:28 <RRSAgent> I have made the request to generate http://www.w3.org/2010/07/30-sparql-minutes.html AlexPassant
17:15:13 <SteveH> SteveH has joined #sparql
18:13:16 <SteveH> SteveH has joined #sparql
18:32:21 <Zakim> Zakim has left #sparql
19:14:31 <AndyS> AndyS has joined #sparql
19:40:16 <SteveH> SteveH has joined #sparql
20:58:18 <AlexPassant> AlexPassant has joined #sparql
22:56:42 <LeeF> LeeF has joined #sparql
23:22:40 <LeeF_> LeeF_ has joined #sparql