RE: [wsrf] ISSUE: Should resource lifetime require WS-Resource.

[Tim Banks <tim_banks@uk.ibm.com>]

I have a problem with some of the language in both Igor's and Rich's
proposals.

Igor said:

> So normatively it may be descibed as follows "Destroy messge
> exchange results in the WS-Resource being destroyed and not
> available for access after that. Any means used previously
> to access the WS-Resource must result in a fault." So that
> is my suggestion.

+1 to the spirit of this (we can describe Destroy without reference to
WS-RP) , but -1 to some of the words.

Firstly, because the target of 'Destroy' may not implement WS-RP (That is
another part of the issue), so it may not  be a WS Resource under the new
terminology.

Secondly,  because " A WS Resource is a Web service through which a
resource can be accessed."...(etc).  So, following Destroy, the Web service
is still available: it responds with a 'ResourceUnknownFault', so we can't
say that the WS Resource is  'destroyed'.   Perhaps we can say  "A Destroy
message exchange results in the resource being destroyed and not available
for access after that. Any means used previously to access the resource
through the [WS Resource/Web service] must result in a fault."


Nor does the EPR become invalid.  Rich said:

> I think there are two possible semantics here, with
> subtle differences between them.
> 1) Stop processing messages sent to this EPR.

-1 to this, because messages sent to the  web service addressed the EPR are
still processed!  The messages simply cause the  response of
'UnknownResourceFault".

> and
> 2) Destroy the underlying resource.
>(snip)
> I think the semantics should be that of
> DestroyResource (#2) and not just
> InvalidateEPR (#1)

+1 to "DestroyResource"   It is the *resource*  which is destroyed.



Regards, Tim Banks
IBM TP Architecture & Technology. Hursley, UK.
Phone: External +44 1962 815639, Internal 245639