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


Help: OASIS Mailing Lists Help | MarkMail Help

bt-models message

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

Subject: Re: F2f meeting

> My recollection is that both were going back in, with some renaming of the
> web-services one. I can't find anything in my notes on this, though they
> (the notes) do say that I was going to write up a use case clarifying an
> out-of-stock example  and/or something on different ways of undoing. I'm
> doing the main document revision first though.

What is the point of putting the Web Services example back in for BTP? It's
in the HP submission specifically to show that multiple one-phase commit
resources shouldn't (IMO) be "coordinated" in a two-phase protocol precisely
because it can lead to the wrong preconceptions about quality of service and
coordination. Multiple one phase aware resources cannot be coordinated in a
reliable manner. As I've mentioned several times, we obviously can't stop
people spoofing BTP by wrapping one-phase resources in a two-phase
interface, and simply not having an undo, but if that happens and they cause
the cohesion to fail, we should report it to the business logic and let the
open market dictate whether that resource is used again.

Granted the Web Services example is the only one that actually mentioned Web
Service in the text. However, that's not a sufficient reason to put it back.
Let's change some of the text in the other examples and explicitly put Web
Service in if that's what bothers people. I think it's kind of implicit


Dr. Mark Little (mark@arjuna.com)
Transactions Architect, HP Arjuna Labs
Phone +44 191 2064538
Fax   +44 191 2064203

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

Powered by eList eXpress LLC