See also: IRC log
<trackbot> Date: 24 November 2009
<mhausenblas> scribenick: Marcelo
<mhausenblas> (as of http://www.w3.org/2001/sw/rdb2rdf/wiki/ScribeList)
<mhausenblas> ACTION-15?
<trackbot> ACTION-15 -- Harry Halpin to make a draft scribelist on wiki -- due 2009-11-24 -- OPEN
<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/actions/15
<mhausenblas> close ACTION-15
<trackbot> ACTION-15 Make a draft scribelist on wiki closed
<mhausenblas> ACTION-16?
<trackbot> ACTION-16 -- Juan Sequeda to init DDL coverage on Wiki -- due 2009-11-24 -- OPEN
<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/actions/16
<mhausenblas> close ACTION-16
<trackbot> ACTION-16 Init DDL coverage on Wiki closed
<mhausenblas> ACTION-17?
<trackbot> ACTION-17 -- Michael Hausenblas to create an issue for vendor-specific data types -- due 2009-11-24 -- OPEN
<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/actions/17
<mhausenblas> ISSUE-2?
<trackbot> ISSUE-2 -- R2RML serializations -- OPEN
<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/issues/2
<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/wiki/Liaisons
<Souri> Eric volunteers to take over RIF and OWL liaisons
<jsequeda> http://www.cs.utexas.edu/~jsequeda/UltrawrapPresentationW3CRDB2RDF.pdf
slide #2
goals of ultrawrap ..
automatically create SPARQL endpoints for legacy relational databases
real-time consistency between relational and RDF
slide #3
<mhausenblas> s/yes, but I need a bit of help/Topic: Ultrawrap presentation
architecture of ultrawrap
<Souri> an alternative: SPARQL querying of relational data
compite time ... create ontology and virtual triple store
run time ... SPARQL to SQL
slide #4
architecture of ultrawrap
slide #5 and #6
first step of the architecture
putative ontology is created automatically from source relational schema
who is taking? how can I know that?
<Souri> ashok
slide #7
<Souri> +q
first-order logic rules transform SQL DDL to OWL
rules cover all possible combination from DDL
slide #8
articles where the approach of ultrawrap is described
<Zakim> ericP, you wanted to (eventually) ask about the differences between RDF and OWL access (SPARQL vs. OWL API?)
Souri: Do you read the data dictionary to get information about DDL?
jsequeda: yes, we do
slide #9
Relational example about students and courses
slide #11
Proof of total coverage of all key combinations
slide #12
All possible combinations of keys and foreign keys are produced (described by a tree)
All these combinations are considered by ultrawrap
Ultrawrap rules are complete in the sense that considered all possible alternatives of keys and foreign keys
This is shown in the DEXA paper mentioned in slide #8
<Zakim> ericP, you wanted to ask if any were degerate forms a la "i don't care if there are 0, 1 or more elts with this identifier"
<Souri> Are you talking about constraint violations in the actual data?
slide #14
the putative ontology is created from the data dictionary
Ahmed: Could you please say a bit more about the rules?
<Souri> +q
slide #15
what is the quality of the putative ontology?
<Souri> -q
If the quality is sufficient, just renaming is needed
If it is not, putative ontology need to be mapped to a domain ontology
slide #16
virtual triple store
represent relational data as triples
virtual triple store uses a table with tree columns: subject, predicate and object
it actually uses a table with five columns (include index values)
slide #18
example of a select statement that output triples
<Souri> +q
thanks: -)
<MacTed> Marcelo - the voice you couldn't recognize was Orri
slide #19
triple view is a union of all the select statements that output triples
slide #20
SPARQL to SQL
slide #21
shows a naive translation of a SPARQL query into SQL
it is a syntactic transformation
this happens in run time
slide #22
SQL optimizer
slide #23
<Souri> -q
shows some examples of query optimization in the system
slide #24
shows the query plans
scribe: triple view plan and optimal plan
<Souri> As I had indicated on 17-Nov-2009 meeting (see minutes) , we'd like very much to present our SQL-based approach on 01-Dec-2009
<mhausenblas> Souri, yes, we'll discuss this in AOB
<Souri> thanks michael
slide #25
summary of what happens in compile and run time
slide #26
shows experimental results
running on Microsoft SQL Server
<Souri> Are these hot (back-to-back) runs?
Test on BSBM on 1 million triples
Comparison was made with Jena
slide #27
Current work
<MacTed> so *most* of those numbers were calculated from published results ... only a couple were actually run
Version 0 of the system was presented (query only the putative ontology)
<mhausenblas> Michael: Ultrawrap is faster than native SQL (Q5 and Q10)?
I lost the phone connection
<mhausenblas> scribenick: mhausenblas
http://www.w3.org/2001/sw/rdb2rdf/wiki/Initial_Round_of_Presentations
Michael: collision OKKAM and Oracle
<Marcelo> I am back :-)
PROPOSAL: Oracle 1 Dec, Eric 8 DEc and OKKAM 15 dec
<Souri> +1
<angela_UNITN> +1
<Marcelo> +1
[adjourned]
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/yes, but I need a bit/Topic: Ultrawrap presentation/ FAILED: s/yes, but I need a bit of help/Topic: Ultrawrap presentation/ Succeeded: s/Ultrawrap presentation of help/Ultrawrap presentation/ Succeeded: s/thanks./Topic: Admin/ Succeeded: s/continues - Michael is a slacker ;)/Topic: Liaisons/ Succeeded: s/Could you please put the URL for the slides on IRC?/Eric volunteers to take over RIF and OWL liaisons/ Found ScribeNick: Marcelo Found ScribeNick: mhausenblas Inferring Scribes: Marcelo, mhausenblas Scribes: Marcelo, mhausenblas ScribeNicks: Marcelo, mhausenblas Default Present: Ashok_Malhotra, +043316876aaaa, +7.791.000.0.aabb, +1.512.471.aacc, [IPcaller], +1.781.273.aadd, +1.603.897.aaee, +1.508.358.aaff, MacTed, whalb, +3539149aagg, jsequeda, Souri, mhausenblas, nunolopes, +39.046.188.aahh, EricP, cygri Present: Ashok_Malhotra +043316876aaaa +7.791.000.0.aabb +1.512.471.aacc [IPcaller] +1.781.273.aadd +1.603.897.aaee +1.508.358.aaff MacTed whalb +3539149aagg jsequeda Souri mhausenblas nunolopes +39.046.188.aahh EricP cygri Regrets: Harry Ben Agenda: http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2009Nov/0056.html Found Date: 24 Nov 2009 Guessing minutes URL: http://www.w3.org/2009/11/24-RDB2RDF-minutes.html People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]