<trackbot> Date: 28 September 2010

hi

ok

<mhausenblas> thanks!

ok Michael

<mhausenblas> scribenick: boris

thanks

PROPOSAL:

topic

admin

admin

accept minutes last meeting

<mhausenblas> PROPOSAL: Accept the minutes of last meeting, see http://www.w3.org/2010/09/21-rdb2rdf-minutes.html

(sorry)

+1

RESOLUTION: WG has accepted the minutes

<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/track/actions/open

OPEN ACTIONS

<mhausenblas> ACTION-72

<mhausenblas> ACTION-72?

<trackbot> ACTION-72 -- Eric Prud'hommeaux to list issues re Datalog approach re PK, FK, etc -- due 2010-09-28 -- OPEN

<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/actions/72

mhausenblas: 72 is done ..
... ericP: are you ok?

<mhausenblas> close ACTION-72

<trackbot> ACTION-72 List issues re Datalog approach re PK, FK, etc closed

mhausenblas: clean up a bit the rest of the actions

<mhausenblas> ACTION: Hausenblas to clean up the actions [recorded in http://www.w3.org/2010/09/28-rdb2rdf-minutes.html#action01]

<trackbot> Created ACTION-73 - Clean up the actions [on Michael Hausenblas - due 2010-10-05].

FPWD "Relational Database to RDF Mapping Language"

Souri: we don't have a draft yet, but working on it
... copy an old SPARQL recommendation and put things there ...
... to discuss sth with richard
... after the discussion today, we have some content there .... html document seems working fine
... should be able to be done this friday, probably something
... if richard has time as well

mihausenblas: what part is technically stuff, and what part is the content?

souri: hhalpin can help with the html

richard: technically side, is hard to get some version control, to all have access to the cvs ...
... content side, we don't have so much yet
....: we need brainstorm things, and discuss

souri: I have some idea what to write ... the content to provide the syntax in turtle
....: then send to richard to take a look

<mhausenblas> Michael: will follow-up with Harry to sort out the technically issues

richard: among the 3 editor have to talk the number of issues for creating the language ...
....: we can use the time now, or schedule sth later

mhausenblas: we can take some time here ... if it won't take so longer
....: use the wiki?

richard: I don't like the idea of using the wiki

<ericP> there's no misery like wiki layout misery

souri: we could you use the telco to discuss the schema and the example
... once we have the cvs we can use it for collaboration

richard: as soon as the cvs works the check-in the things

souri: look at the turtle syntax and xml syntax
... translate xml syntax to turtle syntax and an extended example
... can we discuss that today, and agree for the core content

mhausenblas: what wiki page?

<Souri> http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Schema_in_Turtle

souri: url of the schema

<Souri> http://www.w3.org/2001/sw/rdb2rdf/wiki/Example_of_SQL-based_RDB2RDF_Mapping:_Revision_3

souri: now the url of the example
... extended example
... look at the schema, is very tiny ...
... is it ok or not?
...: ideally to use owl .. we decided not to use

it
....: e.g. cardinality constraint ... this requires owl, so we remove it

richard: can we take later it later? to what extended we need it .... for the definition of the language it doesn't matter

souri: I agree
... to complex otherwise
... later on we can decide

richard: questions .....
... the terminology in the example and the schema, there are some things .... that are inspired in owl more than rdf

souri: examples

richard: for example, class map definition
... rdf level, class is not defined, it is in rdfs
... instance is just a resource

<dan> +q

souri: we use rdfs ....
... in your example you are using classmap

richard: the turtle example just translates things from xml example
... stick to RDF terminology in general
... be aware in which section we need upper levels

souri: worry about the simple idea of classes, seems to clarify things
... not owl, but RDF(S) classes are simple, class and subclass hierarchy
... jump from RDF to RDF(S) is minimal, not to OWL
... RDF(s) is clear to the reader

dan: I agree with that ... this is not rdf or rdf(s) discussion
....: we need to consider when we are going to be successful .... rdf community, database community .... we should be aware .... data quality?

(sorry not hearing very well dan)

mhausenblas: table of the contents?

souri: schema and example
... every aspect of the schema has to be explain
... for each class, what properties, so on
... then some examples and justifications
... some cases at risk, computational risks
... this can be the content right now
... we can improve it later on

richard: two parts .... one part explain what is going here
... second part specific partes
... what is good for, who reads this document

souri: notations, user guide and then some reference

richard: sort of one page of overview that explains the approach of the language ....
....: slides from alex to take this as an example ..... sth of the query and then explain how this is translated in rdf
... after we can go after the individual features, class maps, FK, PK, ..

mhausenlas: what is the doc about? after we read the doc what we can do?

richard: after reading the doc, we will be able to write R2RML docs
... you can't implement the language if you can understand the language

mhausenblas: what we want to provide

richard: specification of the language
... all the knowledge necessary for writing a mapping document
... an implementor has to know all the of things related to the language
... not proposing an implementation
... this is a reference, go to the things one by one
... ideal reader, someone who really wants to learn the language

mhausenblas: we should provide some guidance .... not so much details ...

souri: looking at the sparql doc. ....
... introduction, conventions
... then informative chapter

<Zakim> ericP, you wanted to say that traditionally, w3c writes for implementers

souri: we can follow the same

ericP: job 1 communicate to implementors, answers and test cases
... job 2 users
...: sparql doc pattern in the default mapping

richard: agrees to eric
... one priority: how all it works
... two: be readable to users ... examples, basic
... for now let's focus on the first part
... details of the language
... later on we can explore on how to present the things to general users

<mhausenblas> Michael: made different experience in RDFa case (primer - http://www.w3.org/TR/xhtml-rdfa-primer/ , and syntax - http://www.w3.org/TR/rdfa-syntax/)

richard: for now focus only on specificy later on ...
... on the job 2

mhausenblas agree with the approach
....: not saying to have two separate documents
... each section clearly specifies for whom is

richard: we have to see if we can have two documents or one

<Zakim> juansequeda, you wanted to ask about experience of having different documents instead of one

richard: but later on

<ericP> also, traditionally, w3c relies on o'reilly for user-facing docs, but we've been a bit more merciful about that recently

juansequeda: I don't see the problem of having 2 docs

richard: to souri, the basic building block in all of the different parts of the proposal ... the classmap is the basic block
... about the view map .... the things written there ... I agree with SQL query ... but this not one to one map to class
... dynamically computed classes .... each row map to different class
... what is your opinion of this

souri: we should take that into account ... this is a first draft quickly
...: your mechanism was a view map .....
....: right now is not taking into account ....
... do not change the fundamental aspect, just a small aspect ...
... definition of the schema won't change a lot
... we can think on some extensions of the schema ... once we have the fundamental
... let's get the first working draft, and then we can continue working on that

mhausenblas: continue to this subject, ok?

richard: totally agree with there are a lot of features to discuss
... consider the time, get the basics
... several features after the FWPD
... this view map, can make some fundamental differences in how you think about the language
... important to have an agreement or concensus on this

souri: viewmap collapses into classmaps ...
... use classmaps or viewmaps .... in principle we have an agreement to allow one class or many classes ... many to many rels
... more on syntax ..
... we can discuss on this ...
... it's important ... but let's leave it out for now ... t
... after the FWPD ... we can fix this aspect

mhausenblas: end of week?

souri: yes ... end of week
... focus on completeness and then the details

mhausenblas: pls use the mailing list ...
... to discus things

richard: ... in terms of process ... for now in html

souri: starting working with the sparql doc and then replace things there
... then send mail to the list

richard: working in chapter X, and the other in chapter Y
... use the list?
... email as version system?

<mhausenblas> [adjourned]

I did it

<mhausenblas> trackbot, end telecon

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)

Found ScribeNick: boris
Inferring Scribes: boris
Default Present: mhausenblas, MacTed, +49.153.6.aaaa, boris, juansequeda, Souri, cygri, nunolopes, +1.512.633.aabb, dan, Seema, ericP
Present: mhausenblas MacTed +49.153.6.aaaa boris juansequeda Souri cygri nunolopes +1.512.633.aabb dan Seema ericP
Regrets: Ashok
Agenda: http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2010Sep/0050.html
Found Date: 28 Sep 2010
Guessing minutes URL: http://www.w3.org/2010/09/28-rdb2rdf-minutes.html
People with action items: hausenblas