W3C

- DRAFT -

RDB2RDF

05 Sep 2008

See also: IRC log

Attendees

Present
Andrew_Eisenberg, Ashok_Malhotra, [IBM_Watson], +1.408.504.aabb, orri, +1.937.775.aadd, iv_an_ru, juansequeda, +1.408.504.aaee
Regrets
Chair
Ashok
Scribe
Ashok

Contents


 

 

<iv_an_ru> Hello Ashok

Minutes from July 25 approved.

f2f Oct 30 and 31 in Karlsruhe is confirmed.

Action on Wolfgang to send out meeting logistics

Cancel calls on Sept 26, Oct 3, 17 and 24

<scribe> ACTION: items to Ahmed started email thread [recorded in http://www.w3.org/2008/09/05-RDB2RDF-minutes.html#action01]

Ahmed asks about performance work

Chris Bizer is starting some work in this area

Ashok will forward to XG when he finds out more

Orri: Chris has a SPARQL benchmark

Mapping imposes some penalty but not much

Chris will circulate second revision of SPARQL benchmark within the next 20 days

Will mesaure triple store as well as 2 mapping approaches

Ashok: Are is getting mature

Orri: Certainly for pilot applications

<iv_an_ru> http://virtuoso.openlinksw.com/blog/

Orri has blog with some performance numbers. See above.

Satya: No further progress on the literature survey

Orri will make presentation on performance numbers -- perhaps next week

Satya to remind xxx re. presentation

Satya will follow up with Asio/BBN folks re. presentation

Discussion of deliverables

<iv_an_ru> BTW there should be one language for both RDB2RDF and RDF2RDB :)

<iv_an_ru> RDB-121-RDF

1. Satya/Wofgang will convert wiki to a paper

2. Recommend to the W3C to start a WG to standardize a language for mapping RDB to RDF

Language will not be concerned with whether data is stored as RDF or accessed virtaully

Perhaps a simple/direct mapping version of the language as well as a more fully featured version allowing extensibility

<iv_an_ru> Virtual access requires much more expressivity of the mapping language than just composing RDF dump.

<iv_an_ru> +q

Ivan: We would need a mapping from simple syntax to complex syntax

There is also a performance tradeoff

Orri: We will need the SQL part to be extensible as well. Many systems have proprietary extensions that we would like to exploit

Cathy: If the extensibility string gets very complex it is hard to figure out what's going on
... Does it have to be XML?

Ashok: You start from RDB Schema, right?

Orri: Writing as RDF instance data is very tedious
... In favor human readable syntax rather than expressing parse tree as XML or RDF

Andrew: We could have 2 versions of the language -- human and machine readable and allow mapping between then.

By human readable we mean SPARQL-like not natural language

<iv_an_ru> iv_an_ru owl:instanceOf dbpedia:HomoSapiens ==> I don't want to write XML or RDF, the notation should resemble SQL or a mix of SQL and SPARQL/TURTLE.

<iv_an_ru> At the same time some vendors may decide to export/import descriptions of mappings in RDF form but that's not for initial version of the spec, I guess.

<iv_an_ru> (conference system has disconnected me and do not connect again :(

Orri: I will create initial set of requirements
... I would like some usecases

Can Cathy send a few representative queries

Cathy: I will do that

<iv_an_ru> I'd like some usecases that are benchmarks as well, to ensure that we did not forget something important for performance of RDF Views.

3. Orri's note re. marking updated RDF graphs

Ashok: Not directly in line of the XG

orri: Useful for customizing triple-stores
... Rocommends we wait on this
... The WG that is started should have something to say about this but perhaps the XG need not.

Ashok: So, let's withdraw this from the list of deliverables for the time being.

AOB?

End of XG Telcon

Summary of Action Items

[NEW] ACTION: items to Ahmed started email thread [recorded in http://www.w3.org/2008/09/05-RDB2RDF-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/09/05 16:04:28 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/piloy/pilot/
No ScribeNick specified.  Guessing ScribeNick: Ashok
Inferring Scribes: Ashok

WARNING: No "Topic:" lines found.

Default Present: Andrew_Eisenberg, Ashok_Malhotra, [IBM_Watson], +1.408.504.aabb, orri, +1.937.775.aadd, iv_an_ru, juansequeda, +1.408.504.aaee
Present: Andrew_Eisenberg Ashok_Malhotra [IBM_Watson] +1.408.504.aabb orri +1.937.775.aadd iv_an_ru juansequeda +1.408.504.aaee
Got date from IRC log name: 05 Sep 2008
Guessing minutes URL: http://www.w3.org/2008/09/05-RDB2RDF-minutes.html

WARNING: No person found for ACTION item: items to ahmed started email thread [recorded in http://www.w3.org/2008/09/05-rdb2rdf-minutes.html#action01]

People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report


[End of scribe.perl diagnostic output]