- From: Williams, Stuart <skw@hplb.hpl.hp.com>
- Date: Wed, 17 Apr 2002 16:47:54 +0100
- To: "'www-archive@w3.org'" <www-archive@w3.org>
- Message-ID: <5E13A1874524D411A876006008CD059F192AD3@0-mail-1.hpl.hp.com>
Folks, At the last TBTF telcon I took the action to see what it would take to change the request response MEP description so that it could accomodate temporal overlap between SOAP request and SOAP response. The changes are in sections 6 and 7 of the attached modified version of part 2 (just those sections). The requesting and responding statemachines remain structurally similar to each other and have a simmilar appearance to the previous version. However, the starting state is more clearly marked as are the actions to 'get things moving'. The changes in section 6 should give a good feel for the difference. Only the final state (before success or fail) now waits for transmissions and/or receptions to complete, all other transitions are driven by the the availability of messages to send or the start of messages being received (or the failure of the underlying protocol or a local abort for whatever reason - actually the latter could be a property set in the MEC, but that is mostly editorial). One noticable change is that I split context:CurrentMessage into an InboundMessage and an OutBoundMessage because of the potential for overlap. Most of the 'dense' material in the HTTP binding tables (section 7) is unchanged, although some tables have changed position. The amendments to the binding description were pretty straight forward. So... the quick way to review this is to consider section 6. If that makes sense to you, section 7 should not contain any surprises. (There is a chance I have done this in too much haste... so look out for errors as well). This is up for discussion on the TBTF call on Friday. Best regards Stuart Williams
Attachments
- text/html attachment: OverlappingSRR-MEP_Binding.html
- application/octet-stream attachment: OverlappingRRStd.png
Received on Wednesday, 17 April 2002 11:50:53 UTC