OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

wsbpel message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [wsbpel] Groups - 2003-09-18.BTM.BPEL.syntax.summary.ppt uploaded


David,

thanks.
 but ..

Your suggestion for type makes me uncertain certain we're thinking of
"transaction" in quite the same way.  

I can envisage properties of a transaction, and the need to specify them
( sub-issue is how much is in the bpel proper and how much by
configuration - actually this relates to the general strategic issue of
how much goes in the portable bpel and how much in the perhaps
not-so-portable immediate environment (cf issue 11 discussion)). but I'd
thought of those as being things like "atomic" versus "selective", or
timelimits or "ws-atomic" v "ws-caf(lra)". This possibility is mentioned
in our earlier text input, but I didn't include it in the examples in
the ppt.   Your "type" values (and some of the other fields you suggest)
make me wonder if you are thinking of transaction in the sense of
roughly "a defined multi-party pattern of interaction" - what I think is
considered as "a choreography" in Another Place. I was thinking of
business transaction in sense of an equivalent of a classic ACID
transaction as applied to the (usually) more loosely-coupled world of
web-services - a set of interactions that are subject to a common
decision to complete or reverse (see Alastair's presentation for better
and perhaps more comprehensble descriptions). Both, by history, are
reasonable meanings of "business transaction", but they are really quite
different.

Our proposal was concerned with the ability for bpel processes to
initiate, control, propagate and participate in business transactions
where a supporting generic protocol communicated whether the entities
should complete their work, or parts of it positively (confirmed,
finally completed) or negatively (cancelled, compensated). That could be
a component or aspect of some of independently defined interactions, but
is essential orthogonal to such questions.

It may be I've completely misunderstood and you are talking about
exactly the same kind of beast as I am, but with different names for the
parts. In which case, sorry. (but I really don't understand threadID)

Peter



> -----Original Message-----
> From: David RR Webber [mailto:david@drrw.info] 
> Sent: 24 September 2003 18:44
> To: wsbpel@lists.oasis-open.org; Furniss, Peter
> Subject: Re: [wsbpel] Groups - 
> 2003-09-18.BTM.BPEL.syntax.summary.ppt uploaded
> 
> 
> Peter,
> 
> I've just reviewed your PPT - looks a good start point.  What 
> I'd like to add to the consideration is the following bullet 
> points that add ability to point to transaction properties::
> 
>     - name of transaction
>     - type - XML/EDI/UBL/BOD et al
>     - URL to definition (schema or CAM template)
>     - version
>     - context URL (points to XML that contains context)
>     - threadID (optional - where applicable)
> 
> Thanks, DW.
> 
> > Document Description:
> > Proposed changes to BPEL to support Business Transaction 
> management -
> Peter Furniss & Tony Fletcher presentation
> >
> 
> 


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]