Re: [wsbpel] Issue - 103 - Moving Forward

[Francisco Curbera <curbera@us.ibm.com>]

I feel uncomfortable moving in one direction for XSD variables w/o a clear
idea of how message variables will be dealt with. This may result in
fracturing even more the way we deal with data. I do strongly support
incorporating the BPEL - XPath binding however.

I would also prefer to avoid partially closed issues. Another way to go
might be to open a new "bug" issue ("BPEL specification does not provide
the required XPath 1.0 binding") and use that part of the proposed 103
resolution to close it.

Paco




                                                                       
                      "Yaron Y. Goland"                                
                      <ygoland@bea.com>        To:       Satish Thatte <satisht@microsoft.com>
                                               cc:       Danny van der Rijn <dannyv@tibco.com>, wsbpeltc
                      09/27/2004 08:50          <wsbpel@lists.oasis-open.org>
                      PM                       Subject:  Re: [wsbpel] Issue - 103 - Moving Forward
                      Please respond to                                
                      ygoland                                          
                                                                       




The proposal would specify that the $ syntax would be available for
non-messageType variables only. For the moment I don't intend to
restrict getVariableData for messageType only variables because I hope
we will come up with a resolution for the messageType issue that allows
us to get rid of getVariableData.

The point of having a partial proposal is to let us make progress. We
can agree on the parts we agree on and get them voted on and then use
that as a foundation for moving forward on the issues of messageType and
property access.

             Yaron

Satish Thatte wrote:
>
>
> So if I understand this correctly, you are proposing that we do two
> things with the partial resolution
>
> 1.  Allow BPEL variables of XML types/elements to be bound into XPath
> using the $ notation
>
> 2.  Define the BPEL context for XPath more precisely including
> differentiating normal vs join expressions
>
> I certainly support 2.  I am a bit concerned about a schizophrenic
> resolution of variable binding.  One of the reasons I liked the $
> notation was that it clearly defined the point of binding, eliminating
> any races (at least in my na�ve understanding of XPath) in cases of
> multiple occurrences of getVariableData in an expression, which is
> evaluated concurrently with updates to those variables.  Now we would
> have a split brain on this issue.
>
> Satish
>
> -----Original Message-----
> From: Danny van der Rijn [mailto:dannyv@tibco.com]
> Sent: Monday, September 27, 2004 9:27 AM
> Cc: wsbpeltc
> Subject: Re: [wsbpel] Issue - 103 - Moving Forward
>
> +1
>
> Alex Yiu wrote:
>
>  >
>  > A big +1 from me.
>  >
>  > Regards,
>  > Alex Yiu
>  >
>  >
>  > Yaron Y. Goland wrote:
>  >
>  >> Any time a proposal gets too big it can be difficult to make progress
>  >> because too many issues get mixed up together. As such I propose that
>  >> we break issue 103 into separate proposals, much as we did with issue
>  >> 10.
>  >>
>  >> The first proposal I think should go out for a vote would be a
>  >> modified version of the current 103 proposal that removes the WSDL
>  >> binding to XPATH variables, keeps the getVariableData function and
>  >> doesn't change the current from-specs and to-specs.
>  >>
>  >> This proposal would just include the binding of non-WSDL BPEL
>  >> variables to XPATH variables along with defining the XPATH
>  >> environment in which various BPEL expressions execute.
>  >>
>  >> Then when we settle exactly how we would like to model WSDL variables
>  >> we can decide the future of getVariableData(), how WSDL variables
>  >> will appear in XPATH (including any special forms for doc/lit), if we
>  >> would like to allow properties to be bound to variables, what
>  >> from-spec and to-specs we want to use, etc.
>  >>
>  >> I believe that by breaking this problem up into discrete issues we
>  >> can make more rapid progress.
>  >>
>  >> What does the group think?
>  >>
>  >>     Thanks,
>  >>
>  >>         Yaron
>  >>
>  >> To unsubscribe from this mailing list (and be removed from the roster
>  >> of the OASIS TC), go to
>  >>
>
http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php
.
>
>  >>
>  >>
>  >
>  >
>  >
>  > To unsubscribe from this mailing list (and be removed from the roster
>  > of the OASIS TC), go to
>  >
>
http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php
.
>
>  >
>  >
>  >
>
> To unsubscribe from this mailing list (and be removed from the roster of
> the OASIS TC), go to
>
http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php
.
>
>
>
> To unsubscribe from this mailing list (and be removed from the roster of
> the OASIS TC), go to
>
http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php
.
>
>

To unsubscribe from this mailing list (and be removed from the roster of
the OASIS TC), go to
http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php
.