16:58:30 <RRSAgent> RRSAgent has joined #RDB2RDF
16:58:30 <RRSAgent> logging to http://www.w3.org/2009/12/08-RDB2RDF-irc
16:58:32 <trackbot> RRSAgent, make logs world
16:58:32 <Zakim> Zakim has joined #RDB2RDF
16:58:34 <trackbot> Zakim, this will be 7322733
16:58:34 <Zakim> ok, trackbot; I see SW_RDB2RDF()12:00PM scheduled to start in 2 minutes
16:58:35 <trackbot> Meeting: RDB2RDF Working Group Teleconference
16:58:35 <trackbot> Date: 08 December 2009
16:58:38 <batla> batla has joined #rdb2rdf
16:58:40 <Marcelo> Marcelo has joined #rdb2rdf
16:59:29 <batla> batla has joined #rdb2rdf
16:59:54 <mhausenblas> Zakim, who's here?
16:59:54 <Zakim> SW_RDB2RDF()12:00PM has not yet started, mhausenblas
16:59:55 <Zakim> On IRC I see batla, Marcelo, Zakim, RRSAgent, LeeF, li_ma, Souri, hhalpin, Ahmed, BenSzekely, jsequeda, nunolopes, MacTed, mhausenblas, iv_an_ru, trackbot, ericP
17:00:58 <ericP> Zakim, please dial ericP-office
17:00:58 <Zakim> ok, ericP; the call is being made
17:01:11 <soeren> soeren has joined #RDB2RDF
17:01:11 <mhausenblas> http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2009Dec/
17:02:15 <MacTed> Zakim, who's here?
17:02:15 <Zakim> SW_RDB2RDF()12:00PM has not yet started, MacTed
17:02:16 <Zakim> On IRC I see soeren, batla, Marcelo, Zakim, RRSAgent, LeeF, li_ma, Souri, hhalpin, Ahmed, BenSzekely, jsequeda, nunolopes, MacTed, mhausenblas, iv_an_ru, trackbot, ericP
17:02:23 <MacTed> Zakim, this is 7322733
17:02:23 <Zakim> ok, MacTed; that matches SW_RDB2RDF()12:00PM
17:02:27 <MacTed> Zakim, who's here?
17:02:27 <Zakim> On the phone I see [IPcaller], ??P1, +1.781.907.aaaa, [IPcaller.a], Souri, mhausenblas, +1.562.249.aabb, EricP, OpenLink_Software, +44.131.208.aacc
17:02:30 <Zakim> On IRC I see soeren, batla, Marcelo, Zakim, RRSAgent, LeeF, li_ma, Souri, hhalpin, Ahmed, BenSzekely, jsequeda, nunolopes, MacTed, mhausenblas, iv_an_ru, trackbot, ericP
17:02:33 <Zakim> +[IPcaller.aa]
17:02:43 <hhalpin> Zakim, aa is hhalpin
17:02:43 <Zakim> sorry, hhalpin, I do not recognize a party named 'aa'
17:02:44 <MacTed> Zakim, OpenLink_Software is temporarily MacTed
17:02:44 <Zakim> +MacTed; got it
17:02:47 <MacTed> Zakim, mute me
17:02:47 <Zakim> MacTed should now be muted
17:02:52 <hhalpin> Zakim, ??aa is hhalpin
17:02:52 <Zakim> sorry, hhalpin, I do not recognize a party named '??aa'
17:02:58 <nunolopes> Zakim, nunolopes is with mhausenblas
17:02:58 <Zakim> +nunolopes; got it
17:03:22 <mhausenblas> batla are you scribing?
17:03:37 <hhalpin> If no one else can do it, I can scribe.
17:04:03 <cygri> cygri has joined #rdb2rdf
17:04:07 <hhalpin> Zakim, mute me
17:04:07 <Zakim> sorry, hhalpin, I do not know which phone connection belongs to you
17:04:16 <soeren> I can not dial in today, but will observe proceedings here via IRC
17:05:02 <hhalpin> Zakim, [IP call.aa] is hhalpin
17:05:02 <Zakim> I don't understand '[IP call.aa] is hhalpin', hhalpin
17:05:10 <hhalpin> Zakim, [IP caller.aa] is hhalpin
17:05:10 <Zakim> I don't understand '[IP caller.aa] is hhalpin', hhalpin
17:05:16 <Zakim> +jsequeda
17:05:30 <Zakim> +Seema
17:06:38 <mhausenblas> scribenick: batla 
17:06:39 <Seema> Seema has joined #RDB2RDF
17:06:44 <mhausenblas> Agenda: http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2009Dec/0005.html
17:06:54 <mhausenblas> Topic: Admin
17:07:06 <batla> batla will scribe
17:07:59 <Souri> do you have the url?
17:08:00 <jsequeda> it's this one, right? http://www.w3.org/2009/Talks/1215-SWObjects-egp/#(1)
17:08:45 <mhausenblas> RESOLVED: http://www.w3.org/2009/12/01-RDB2RDF-minutes.html accepted
17:08:56 <mhausenblas> Topic: Presentation of SWObjects (Eric Prud'hommeaux)
17:09:16 <mhausenblas> http://www.w3.org/2009/Talks/1215-SWObjects-egp
17:09:36 <hhalpin> ericP: this is work on SWObjects with HCLS IG
17:09:45 <hhalpin> ... main problem is sociopathic names in RDF
17:09:48 <hhalpin> ... so we have two graphs
17:09:51 <hhalpin> ... a stem graph
17:10:00 <hhalpin> ... and an interface graph that is actually queried.
17:10:18 <hhalpin> ... 3 approaches
17:10:31 <hhalpin> ... 1) We use DDL to create a triple view...
17:10:37 <MacTed> (slide 4 typo -- s/preciate/predicate/)
17:10:49 <hhalpin> 2) RDF Schema
17:10:57 <hhalpin> 3) Normative RDF Representation
17:11:30 <hhalpin> So what we do is use a stem graph representation and then create using SPARQL constructs we create representation data.
17:11:36 <hhalpin> Table and columns to types of predicates
17:11:47 <hhalpin> SQL to create interface views vs. SPARQL to create interface views.
17:12:23 <hhalpin> Goal: What we want is clear algebra for the production of a stem graph, and a clear algebra for a SPARQL to SQL translation, and a clear algebra from the SQL results to RDF notes.
17:12:29 <hhalpin> s/notes/nodes
17:13:02 <hhalpin> ericP: Slide 6: SPARQL to SQL results
17:13:13 <hhalpin> ... a set of algebra functions
17:13:19 <hhalpin> ... a SQL algebra subset
17:13:41 <hhalpin> ... here is the subet of SQL that we can use for SPARQL.
17:14:17 <hhalpin> ... there are some issues, i.e. in particular when we are going to different data sources, how to express that in SQL?
17:14:39 <hhalpin> ... ASK and CONSTRUCT are a bit tricky, is ASK exists? What is the name of it, what is its relationship?
17:14:48 <hhalpin> ... we could easily do it, is it useful?
17:15:20 <hhalpin> ... Slide 7: Here is an algebra for the stem mapping
17:15:36 <hhalpin> ... the stem mapping is called that due to the fact that you have to provide a stem URI
17:15:45 <hhalpin> ... we need to create a stem graph that meets the same use cases as relational data
17:15:54 <hhalpin> ... if you view t he log
17:16:13 <hhalpin> ... what queries could be easily transformed?
17:16:35 <hhalpin> DanMiranker: Would like to investigate these mappings in detail
17:17:20 <hhalpin> Orri: If you are part of SPARQL 1.0 you cant transform everything, but with 1.1 features like aggregate, it should be possible to map all of SQL.
17:17:38 <hhalpin> ericP: Lets try to do 1.0 first, and then we can do the later versions later.
17:18:07 <hhalpin> ericP: Slide 8 how do we take this relational data and map it coherently
17:18:11 <hhalpin> ... can we get the naming right
17:18:21 <hhalpin> ... re linked data principles
17:18:59 <hhalpin> ... the same answers received by SPARQL should match SQL
17:19:05 <hhalpin> ... this algebra is very typical
17:19:14 <hhalpin> ... for any query language, projections and graph pattern s
17:19:24 <hhalpin> ... union, disjunction, conjunctions, and optionals
17:19:32 <hhalpin> ... unlimted graph patterns
17:20:18 <hhalpin> ... majority of work is the same as queries over a triple view of SQL
17:21:01 <hhalpin> ... basic premise: table algebra is a set of constraints coupled with a subject, that identifies a table alias. 
17:21:40 <hhalpin> ... .I have done this in pel a while ago
17:21:48 <hhalpin> s/pel/perl
17:21:55 <hhalpin> ericP: Now in SQL
17:22:04 <hhalpin> .. slide 9
17:22:19 <hhalpin> ericP: the way you handle the graph patterns are pretty consistent
17:22:31 <hhalpin> ... what is in algebra?
17:22:37 <hhalpin> ... bgp mapping
17:22:44 <hhalpin> ... mapping between sparql and sql
17:24:01 <hhalpin> ... big difference is whether or not for triple stores we add a new table for new graph patterns
17:24:11 <hhalpin> ... null guards are completely different in sparql will get to that later
17:24:20 <hhalpin> ... in my implementation, I disallowed variable predicates
17:24:25 <hhalpin> ... I think its doable
17:25:15 <hhalpin> ... but if you disallow wildcard predicates you get something you can implement the SPARQL in about the same time as SQL re compile time and identical time for query.
17:25:21 <hhalpin> ... runtime
17:25:52 <hhalpin> Orri: Allowing variables in SPARQL, we allow them in Virtuoso
17:26:05 <hhalpin> ... but we extend SQL in order to avoid complicated unions
17:26:24 <hhalpin> ... however, what ericP said is true, that without that we can get more or less same overhead
17:26:30 <hhalpin> ... we can compile complex unions
17:26:45 <hhalpin> ... but if your system knows its looking at virtuoso
17:27:07 <hhalpin> ... then we can define the mapping in terms of algebra we can now we have done in correctly.
17:27:26 <hhalpin> ... conjunction is going to be the same
17:27:31 <hhalpin> ... but nulls are different
17:27:37 <hhalpin> ... a null in SPARQL is a missing triple
17:27:50 <hhalpin> ... in SQL, its a special value thats not equal to itself.
17:28:11 <hhalpin> ... curlies inside of query.
17:28:19 <hhalpin> ... I wrote one down in terms if graph constraints
17:28:32 <hhalpin> ... since I want to eventually hint at what we could do with query constraints.
17:28:40 <hhalpin> Orri: You know that maps to unbound in SQL
17:28:53 <hhalpin> ericP: When I do a query for a first name and last name
17:29:05 <hhalpin> ... got a value for that
17:29:28 <hhalpin> ... sql you would not get a result
17:29:33 <hhalpin> ... sparql you would get an answer
17:30:10 <hhalpin> orri: unbound is equal to anything, null is equal to nothing
17:30:18 <hhalpin> ericP: Slide 12
17:30:47 <hhalpin> ... we want to make sure with options that the entire optional graph is transparent
17:30:57 <hhalpin> ... the temptation is to a left outer join on just one nested table
17:31:12 <hhalpin> ... but really you want to get all the tables joined.
17:31:34 <hhalpin> ... otherwise you get cardinality problems (see my paper)
17:31:43 <hhalpin> ... the other problem is leading optionals in SPARQL
17:31:59 <hhalpin> ... so we have to do a bit of trickery
17:32:09 <hhalpin> ... i.e. a one column table
17:32:21 <hhalpin> ... just make sure upstream processing throws that column away
17:32:34 <hhalpin> Slide 13: Things that are unbound in SPARQL
17:32:49 <hhalpin> ericP: again, we have to do a little trick
17:33:15 <hhalpin> ... a SQL query executes the same as a SPARQL query, a potentially null attribute
17:33:29 <hhalpin> ... have to make it explicit that it is not null
17:33:34 <hhalpin> ... its just tedious but doable
17:33:40 <hhalpin> ... Slide 14
17:33:49 <hhalpin> ... all the algebraic operators in SPARQL
17:34:04 <hhalpin> ... like string URI all of these can be handled by schema if you have RDF schema
17:34:25 <hhalpin> ... data types we can just handle ISO ones
17:34:55 <hhalpin> ... when I asked for somebodys name, we dont use xsd:strings
17:35:17 <hhalpin> ... which way is world better  lang tags or data types from XSD
17:35:34 <hhalpin> ... now that we can query this kinda sociopathic representation of a RDB store
17:35:51 <hhalpin> ...  how can we allow access using commonly known URIs like foaf:name
17:36:14 <hhalpin> ... nice thing is interface graph can entirely change shape
17:36:23 <hhalpin> ... can do anything, extra joins etc.
17:36:35 <hhalpin> ... as long as we can map it to stem graph via SPARQL construct
17:36:48 <hhalpin> ... magic stems
17:37:27 <hhalpin> ... creates virtual tables
17:37:40 <hhalpin> ... that can mix with concrete tables and result in more virtual tables
17:37:49 <hhalpin> ... but we can eventually resolve it to queries over concrete tables
17:37:59 <hhalpin> ... same approach re concrete constructs 
17:38:09 <Souri> q+
17:38:22 <jsequeda> q+
17:38:35 <hhalpin> ... push query over virtual interface table to a query that can then be calculated over sociopathic original representation
17:38:50 <hhalpin> Souri: We have query rewriting systems in Oracle, very important technique
17:39:04 <hhalpin> ericP: Similar to goal flattening in rules
17:39:43 <hhalpin> DanielMiranker: No use of magic sets 
17:39:48 <mhausenblas> ack jsequeda 
17:39:53 <hhalpin> DanielMiranker: in literature yet
17:40:05 <hhalpin> ... we use this in Ultrawrap
17:41:21 <hhalpin> Souri: We support this, but just differently
17:41:33 <Souri> That was not me
17:41:43 <mhausenblas> ack Souri 
17:41:55 <hhalpin> s/Souri/Orri
17:42:08 <hhalpin> Souri: What is the power of that approach as regards mapping SQL
17:42:14 <hhalpin> ... for example, with analytical functions
17:42:27 <hhalpin> ... how could that be done with SPARQL, could this part of SQL be expresssed?
17:42:45 <hhalpin> ericP: My goal was to stick to SPARQL 1.0 and eventually SPARQL 1.1
17:43:22 <hhalpin> Souri: Possible to transmit that as regards rewriting it over SQL, we can map that easy, the main issue is expressive power
17:43:38 <hhalpin> ... I am not interested in SQL tables as a RDF consumer
17:43:56 <hhalpin> ... so analytical funtions would just get some overall aggregates that we could look at as RDF
17:44:04 <hhalpin> ... and we would just query that view looking at SQL view definitions
17:45:24 <hhalpin> Souri: We can aggregate relational views that are results of analytical f unctions
17:46:26 <hhalpin> ... the main difference between our approach and your approach
17:46:26 <hhalpin> ... is that your body is the SPARQL query
17:46:40 <hhalpin> ... again, my main question is power
17:47:10 <hhalpin> ... we want to look at view definition is completely in SQL; and then on top we map
17:47:17 <hhalpin> ... just to map the columns into predicates
17:47:25 <hhalpin> ... and defining some constraints
17:47:35 <hhalpin> ... the main difference is the expressive power of SQL vs. SPARQL 
17:47:40 <hhalpin> ... for creating the views
17:48:01 <hhalpin> Orri: Minus a few small differences, we can make them more or less equivalent
17:48:09 <hhalpin> Souri: Is it SPARQL we are talking about?
17:48:16 <hhalpin> ... or relational data?
17:49:11 <hhalpin> Ahmed: If you do the work in SQL that is easier, and if you do it outside you need to be careful
17:49:26 <hhalpin> ericP: But this approach should work over anything you can express in SQL
17:50:03 <hhalpin> ... we have to be careful with arbitrary number of extensions
17:50:18 <hhalpin> ... my approach hopes to be portable
17:50:24 <hhalpin> ... spreadsheets, nlp, etc.
17:51:52 <hhalpin> Souri: Read only approach, we can only create views on some kinds of tables, 
17:51:52 <hhalpin> ... when we are talking about spreadsheets and etc.
17:51:52 <hhalpin> ... we are not as deep as part of relational data.
17:51:52 <hhalpin> ... the complexity of the domains are different, so I was thinking RDB2RDF should take into account complexity of relational data.
17:51:55 <hhalpin> ericP: performance
17:52:43 <hhalpin> ... LLR(1) with SPARQL and SQL 
17:52:55 <hhalpin> ... slide 17
17:52:59 <hhalpin> integrtiy constraints
17:53:42 <hhalpin> ... we can push them into sql into two ways
17:53:52 <hhalpin> ... every column that was mentioned in antecedent graph
17:54:08 <hhalpin> ... slide 18
17:54:21 <Ahmed> I hav ea conference call @10:00am sharp, I will split @9:59 - thanks.
17:54:24 <hhalpin> ... we may want to put an antecedent contruct on everything in our attributes
17:55:09 <hhalpin> ... (goes through example) 
17:55:34 <hhalpin> slide 19
17:55:41 <hhalpin> ericP: Integrity constraint options
17:55:44 <hhalpin> ... anything with a constraint
17:55:57 <hhalpin> ... shortest path through all those variables, the magic set view for graph connections
17:56:05 <hhalpin> ... all the paths for everything that is connected
17:56:13 <hhalpin> ... I have used syntactic hints
17:56:20 <hhalpin> ... had no integrity constraints
17:57:11 <hhalpin> ... it didnt matter what it was coming from
17:57:22 <hhalpin> ... so we did virtual views of constructs and role flattening
17:57:37 <hhalpin> ... slide 21
17:57:42 <hhalpin> ... hcls requirements
17:59:08 <hhalpin> ... sparql 1.0 term rewriting, everything is done in interface graph...
17:59:08 <hhalpin> ... rdf construct
17:59:08 <hhalpin> ... turn uris of one form into another re regular expressions of named patterns
17:59:08 <hhalpin> ... come through membership, see wiki for more details
17:59:28 <Zakim> - +1.781.907.aaaa
17:59:34 <hhalpin> ericP: thats it for me.
18:00:10 <hhalpin> DanielMiranker: We are interested in this algebra, since we wanted to compile DDL as well
18:00:18 <hhalpin> ... we did that all that of DataLog
18:00:25 <hhalpin> ... looking at Datalog to relational algebra work.
18:00:42 <Zakim> -??P1
18:00:49 <Zakim> -[IPcaller.a]
18:00:50 <hhalpin> ericP: Would that handle everything like leading optionals, i.e. in Datalog
18:00:59 <mhausenblas> RRSAgent, draft minutes
18:00:59 <RRSAgent> I have made the request to generate http://www.w3.org/2009/12/08-RDB2RDF-minutes.html mhausenblas
18:01:00 <hhalpin> DanielMiranker: Yes, I think it would
18:01:10 <mhausenblas> RRSAgent, make logs public
18:01:12 <hhalpin> Meeting Adjourned
18:01:13 <Zakim> -[IPcaller.aa]
18:01:41 <mhausenblas> Zakim, list attendees 
18:01:41 <Zakim> As of this point the attendees have been [IPcaller], +1.781.907.aaaa, mhausenblas, Souri, +1.562.249.aabb, EricP, +44.131.208.aacc, MacTed, nunolopes, jsequeda, Seema
18:01:51 <mhausenblas> RRSAgent, draft minutes
18:01:51 <RRSAgent> I have made the request to generate http://www.w3.org/2009/12/08-RDB2RDF-minutes.html mhausenblas
18:01:52 <Zakim> -Souri
18:01:53 <hhalpin> hopes to see this conversation keep up on the listserv!
18:01:55 <hhalpin> take care!
18:01:57 <Zakim> -jsequeda
18:01:59 <Zakim> -Seema
18:02:01 <Zakim> - +44.131.208.aacc
18:02:02 <Zakim> -mhausenblas
18:02:02 <Zakim> -MacTed
18:02:07 <Zakim> - +1.562.249.aabb
18:02:14 <Marcelo> Marcelo has left #rdb2rdf
18:03:13 <Zakim> -[IPcaller]
18:29:42 <hhalpin> hhalpin has left #rdb2rdf
19:01:19 <cygri> cygri has joined #rdb2rdf
19:01:28 <cygri_> cygri_ has joined #rdb2rdf
19:05:00 <Zakim> disconnecting the lone participant, EricP, in SW_RDB2RDF()12:00PM
19:05:02 <Zakim> SW_RDB2RDF()12:00PM has ended
19:05:04 <Zakim> Attendees were [IPcaller], +1.781.907.aaaa, mhausenblas, Souri, +1.562.249.aabb, EricP, +44.131.208.aacc, MacTed, nunolopes, jsequeda, Seema
19:34:43 <nunolopes> nunolopes has joined #RDB2RDF
20:28:13 <Zakim> Zakim has left #RDB2RDF
20:58:29 <nunolopes> nunolopes has joined #RDB2RDF
21:02:28 <jsequeda> jsequeda has joined #rdb2rdf
21:37:15 <cygri_> cygri_ has joined #rdb2rdf
22:36:13 <LeeF> LeeF has joined #rdb2rdf
22:57:27 <cygri> cygri has joined #rdb2rdf
23:03:43 <jsequeda> jsequeda has joined #rdb2rdf
23:14:52 <cygri_> cygri_ has joined #rdb2rdf