RDB2RDF Working Group Teleconference -- 23 Nov 2010

<trackbot> Date: 23 November 2010

<mhausenblas> scribenick: MacTed

<boris> I had it in a previous telco ...

Admin

<mhausenblas> PROPOSAL: Accept the minutes of last meeting, see http://www.w3.org/2010/11/16-rdb2rdf-minutes.html

<juansequeda> +1

RESOLUTION: Accept the minutes of last meeting, see http://www.w3.org/2010/11/16-rdb2rdf-minutes.html

<hhalpin> +1

Harry loves RDB2RDF, but has too few hours to do that and his new responsibilities. Ivan Herman will be picking up Harry's admin roles.

Review of R2RML Test Cases

:-)

<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases

mhausenblas: having created mapping syntax, now need to test to confirm it delivers as designed

<mhausenblas> Michael: let's start with the first two

<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases#R2RMLTC0000

Boris: First case, obviously very simple...

<boris> http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases#R2RMLTC0000

<hhalpin> +1 on new version of R2RML

MacTed: Test Case #0 appears to lack class, attribute definitions?

Souri: R2RML mapping should only cover tables with data. Empty tables should be left to Direct Mapping.

hhalpin: since Direct Mapping already includes databases & code, best to re-use those samples

<hhalpin> particularly for direct mapping

mhausenblas: proposes to put test case #0 on hold, given active discussion

<mhausenblas> PROPOSAL: put TC0000 on hold depending on Direct Mapping as default

<hhalpin> notes we can publish while that is on hold.

<boris> The purpose of the R2RML test cases is to verify and validate the specification of the R2RML language. Moreover, these test cases can be useful to create a test harness or automated test framework for testing the interoperable implementors, i.e. R2RML engines.

<Souri> We will allow an RDFS class to be generated (e.g., exa:emp rdf:type rdfs:Class) even when the logical table identified in a TriplesMap is actually empty.

mhausenblas: discussion around what we're trying to do, what these test cases are supposed to be testing ...

<hhalpin> +1

<boris> +1

RESOLUTION: put TC0000 on hold depending on Direct Mapping as default

<Souri> +1

<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases#R2RMLTC0001

Boris: TC0001 - 1 table, 1 single column, 1 single row ...
... only RDB instance data is made into RDF; no schema info; no classes, etc.

<ericP> am i invited to contribute direct mapping tests here as well?

<ericP> therer are a bunch in the document, but some more i have in my implementation tests

<juansequeda> ericP, don't you think that all these test cases should have two ouputs: the R2RML customized mapping and the Direct Mapping one?

<ericP> juansequeda, i thought i just heard that these tests shouldn't have an empty r2rml configuration

<Marcelo> I agree with juansequeda

<juansequeda> ericP, each test case has "Expected Result"

<boris> yes, ericP

<juansequeda> it should also have "Expected Default Mapping Result"

<ericP> how about if an empty r2rml be a direct mapping test?

<juansequeda> and take the same input schema/data

<ericP> that sounds good to me

<ericP> sure

<ericP> sorry, on two telecons

<ericP> (only one mute button)

<ericP> sure

<mhausenblas> ACTION: Juan and Eric to draft Direct Mapping TC [recorded in http://www.w3.org/2010/11/23-rdb2rdf-minutes.html#action01]

<trackbot> Created ACTION-81 - And Eric to draft Direct Mapping TC [on Juan Sequeda - due 2010-11-30].

<hhalpin> I'd prefer these all to be part of the same test-case document

<hhalpin> just to keep life simple

+1 hhalpin

scribe: discussion about need for Direct Mapping test cases, and potentially integrating all test cases into a single document ...

ericP: sees a lot of framework that will be repeated across Direct Mapping & R2RML
... the pain of editing will be justified by the end result

mhausenblas: needs to be clear, this is the DM stuff and this is the R2RML stuff...
... objections to TC0001?

group ... confusion stemming from labeling, wording of descriptions...

<mhausenblas> remove rr:datatype xsd:string

<mhausenblas> PROPOSAL: Accept TC0001 and remove the rr:datatype

<Souri> +1 (modulo the rr:tableOwner inclusion at some point)

RESOLUTION: Accept TC0001 and remove the rr:datatype

<boris> I'll inlcude the rr:tableOwner in the whole test cases

<hhalpin> move review to list-serv?

<hhalpin> try to get everyone to do a review of it before next telecon?

<boris> zakin, mute me

Implementation

<mhausenblas> PROPOSAL: Create a Wiki page with status of implementations (and plans thereof)

Souri: R2RML doc has been much changed from previous, much improved

<Souri> -(-1) => no objections

RESOLUTION: Create a Wiki page with status of implementations (and plans thereof)

<mhausenblas> ACTION: Hausenblas to create Wiki page with implementation status (and plans thereof) [recorded in http://www.w3.org/2010/11/23-rdb2rdf-minutes.html#action02]

<trackbot> Created ACTION-82 - Create Wiki page with implementation status (and plans thereof) [on Michael Hausenblas - due 2010-11-30].

R2RML status

<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/r2rml/

<ericP> boris, could you review what i've put in http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases#R2RMLTC0005b ?

Souri: week after FPWD was published, reworked into much simpler structure, while test cases being developed
... would like in-depth critical review as soon as possible, to replace FPWD with 2ndPWD

<ericP> boris, i didn't renumber after http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases#R2RMLTC0005b (note the 'b' at the end)

<ericP> we may want names which don't imply an order

<mhausenblas> ACTION: Ashok to review R2RML ED [recorded in http://www.w3.org/2010/11/23-rdb2rdf-minutes.html#action03]

<trackbot> Created ACTION-83 - Review R2RML ED [on Ashok Malhotra - due 2010-11-30].

<juansequeda> ericP, does the 'b' imply direct mapping?

<ericP> juansequeda, it's just 'cause i was lazy and didn't want to renumber 6, 7, 8...

<ericP> i stuck it after 5 'cause it was closest in concept to 00005

<juansequeda> ericP, ok, just confirming

<ericP> ahh, 'b' is for "bis", how's that excuse?

<mhausenblas> LeeF, could you review R2RML as well?

<mhausenblas> ACTION: Hausenblas to ask Lee if he can review R2RML ED [recorded in http://www.w3.org/2010/11/23-rdb2rdf-minutes.html#action04]

<trackbot> Created ACTION-84 - Ask Lee if he can review R2RML ED [on Michael Hausenblas - due 2010-11-30].

<ericP> boris, i think were wondering about the "discriminating query"

<Souri> latest R2RML Editor's draft => http://www.w3.org/2001/sw/rdb2rdf/r2rml/Overview.html

<ericP> it was intended to record a query which would reveal whether you have the correct behavior

<ericP> in this case, an aggregate would reveal whether you have preserved cardinality

<mhausenblas> Michael: Ted please look at http://www.w3.org/2001/sw/rdb2rdf/status/2010/

<ericP> i'll add the SPARQL query as well, but feel free to push back on this

<boris> ericP ... I see ...

<boris> but there is no time

<boris> now

Direct Mapping Status

Marcelo: have been working on DATALOG part of document with juansequeda
... particularly on how to handle foreign keys
... more-or-less know how to do it, need to write it down

juansequeda: there are some noted issues, not sure when/how to start addressing those

mhausenblas: you three editors are drivers of the document; put things forward for agenda, discussion will happen...
... agenda additions *can* also happen in the moment

ericP: feels like this doc has drifted a lot from original WG resolution, wonders whether it needs a full group revisit

<hhalpin> expandability plus collapsable buttons seems minor to me

ericP: rearrangement of headers; expand/collapse buttons a questionable thing

<mhausenblas> Michael: it should not be empty; English text should be the default (non-hideable)

<ericP> http://www.w3.org/2001/sw/rdb2rdf/directGraph/#iris-details1

<hhalpin> here's official resolution: RESOLUTION: http://www.w3.org/2001/sw/rdb2rdf/directMapping/ with Juan, Marcelo and Eric as editors based on Richard's proposal as of http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2010Nov/0052.html and try to work in J&M's IRI and Triple generations part; move hierarchical table and the M-M mappings into Ed note; datalog as a separate section; Eric perform merge with review/approval/consensus of Juan, Marcelo, & Eric

<hhalpin> from http://www.w3.org/2010/11/09-rdb2rdf-minutes.html

<mhausenblas> ACTION: Ted to review Direct Mapping ED [recorded in http://www.w3.org/2010/11/23-rdb2rdf-minutes.html#action05]

<trackbot> Created ACTION-85 - Review Direct Mapping ED [on Ted Thibodeau - due 2010-11-30].

<mhausenblas> ISSUE-5?

<trackbot> ISSUE-5 -- Default vs. direct mapping -- open

<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/issues/5

MacTed: if there are to be *any* collapsible headers, I think they should *all* be so

<mhausenblas> [adjourned]

mhausenblas -- should tell trackbot to end meeting, yes?

<mhausenblas> pls

trackbot, end meeting

hello, trackbot.....

<mhausenblas> trackbot, stop meeting

<trackbot> Sorry, mhausenblas, I don't understand 'trackbot, stop meeting'. Please refer to http://www.w3.org/2005/06/tracker/irc for help

or maybe it can't hear the scribenick?

<scribe> scribenick:

<MacTed> trackbot, end meeting

<mhausenblas> trackbot, end telecon

- DRAFT -

RDB2RDF Working Group Teleconference

23 Nov 2010

Attendees

Contents

Admin

Review of R2RML Test Cases

Implementation

R2RML status

Direct Mapping Status

Summary of Action Items

Scribe.perl diagnostic output