Databinding WG Teleconference

 

 

<trackbot> Date: 13 December 2005

<dino> trackbot can do all the setup for you, but I didn't know the bridge data yet

Meeting Administrivia

two f2f meeting on their way

F2F Meetings

Having trouble setting the date and location .. the meeting might be end of January in sophia Antipolis, followed by the technical plenary in Mandeileu (same area)

--

Status of Deliverables

Roadmap document

the first issue from SAP is targeting partly the roadmap

the roadmap is currently the basic pattern and advanced patterns documents

it would be good to have the test suite and example developped as close as possible to the specifications

also ways to identify patterns

ISSUE-1: http://www.w3.org/2005/06/tracker/databinding/issues/1

ISSUE-1: Scope of Structures to be Addressed

pauld: we will have a roadmap document that will contain what you suggest, is that what you had in mind?

vlad: I didn't see the roadmap document, so that why my thoughts on it

pauld: is the issue also to define rationale to appear in our Recs?

yves: having item4 of issue 1 has a separate issue will make it more technical than just roadmap

pauld: about item3, what do you mean by "uncommon" datastructures?

vlad: in our language we are using structures that won't be covered, there will be no patterns, but there should be a way to map things to be language neutral

sekhar: there should be a way to customize your representation, but in the end you have the issue of the representation available in your end-language
... xml schema is there to help create things based on defined type
... for example we can use annotation in the data representation

pauld: vlad's comment are really code driven

vlad: it's about hints to reconstruct structures

pauld: sekhar made it clear that it is dangerous to map too closely the internal structure to XML
... I'll put item3 as a separate issue
... about item2, why did you call out those structures?

things like pointers looks like trees in soap sec5, wonder if the group was interested in that or not

scribe: defer to the list

it would be good to have a list of things we do.

and have people starting to send one pattern per issue (like vector)

Issue-2 is about the test suite

pauld: (speaking on behalf of jon)

the TS should have tests which may be presented as WSDL so that we can test directly in existing toolkits

Sekhar: would like a clarification. Some vendors have test suites or use existing test suite (like W3C's XML schema)

not sure about the value of this new test suite

pauld: as part of the W3C process, test suite help demonstrating interop

one way is to define an echo service that will just output back the wanted structure

also having a TS helps testing specific points related to that specification

sekhar: databinding vendors are doing better products relative to one specific language, for example toolkit that target java. there are concept that don't map well to programming language (like restrictions)

pauld: there is a difference bewteen "one vendor is doing a better job" and how to model things to maximize the users' experience

regardless of the vendor used

sekhar: if you have complex type derived by extension, it will map nicely, it will bind nicely to programming languages, but if you start with a schema only construct it might not map well to programming languages

pauld: we are chartered to describe datastructures, but we can document alignement (with programming languages)
... it would be good to have discussion starting on the list, about what structures we will define and the alignment with programming languages

any AOB?

pauld: I'll set the date for the f2f, please answer the questionnaire

http://www.w3.org/2002/09/wbs/38435/FirstF2F/

ADJOURNED