W3C

XML Protocols Telcon of 12 April 2006

12 Apr 2006

Agenda

See also: IRC log

Attendees

Present
BEA Systems, David Orchard (DaveO)
IBM, Chris Ferris (Chris, Chris)
IBM, Noah Mendelsohn
Nokia, Mike Mahan
Sonic, Glen Daniels
Sun Microsystems, Marc Hadley
Sun Microsystems, Pete Wenzel
Tibco, David Hull (dhull)
W3C, Yves Lafon
Regrets
Oracle, Anish Karmarkar
Absent
Iona Technologies,
Sonoa Systems, Vikas Deolaliker
Excused
Ericsson; Nilo Mitra
Oracle, Jeff Mischkinsky
Chair
Mike Mahan
Scribe
Yves



Minutes approval

march 4th minutes approved

april 6th minutes approved

Action Items

Chris: pending

Comments

Mike: Mark Baker's comment at http://lists.w3.org/Archives/Public/xmlp-comments/2006Mar/0000.html

Noah sent a reply to Mark

<scribe> ACTION: Yves to open an issue and send closing email immediately [recorded in http://www.w3.org/2006/04/12-xmlprotocol-minutes.html#action01]

ROR

Mike: As Dave and Anish are out, will be discussed later

SC2

http://lists.w3.org/Archives/Public/xml-dist-app/2006Mar/0051.html

Noah: did we discuss the fact that a request has an envelope?

Chris: the issue is that we are not precise enough wrt SOAP message and SOAP envelope

is an HTTP GET a SOAP message?

Noah: yes
... we should clarify the definition of "message" in part1, or be more precise in the bindings like "the request MUST contain an envelope"

<Noah> The request message MUST contain a SOAP envelope.

<Noah> The response MAY contain a SOAP envelope, or else the response must include a binding-specific indication that the request has been received.

<DaveO> I like that.

<Noah> Thanks.

<Chris> sorry, I thought I was connected

<Chris> could someone cut-n-paste Noah's proposed text?

<Noah> The response MAY contain a SOAP envelope, or else the response must include a binding-specific indication that the request has been received.

DaveO: if there is an envelope back it may also be a binding-specific indication that the request has been received.

<Noah> The response MAY contain a SOAP envelope, and MUST in all cases include a binding-specific indication that the request has been received.

<Noah> The rec sez:

<Noah> The SOAP Request-Response MEP defines a pattern for the exchange of a SOAP message acting as a request followed by a SOAP message acting as a response. In the absence of failure in the underlying protocol, this MEP consists of exactly two SOAP messages.

<Chris> what if we changed " indication that the request has been received." to "indication as to the disposition of the request."

<DaveO> I'm liking the MAY, and MUST formulation..

<DaveO> chris agrees to dispose of the dispose suggestion

<Mike> The request message MUST contain a SOAP envelope.

<Mike> The response MAY contain a SOAP envelope, and MUST in all cases include a binding-specific indication that the request has been received.

<Noah> bye DaveO

<Mike> In the absence of failure in the underlying protocol, this MEP consists of exactly two messages.

<Noah> Status quo text for response only:

<Noah> " In the absence of failure in the underlying protocol, this MEP consists of exactly two messages, only one of which is a SOAP message:"

<Noah> Therefore, for ROR:

<Noah> In the absence of failure in the underlying protocol, this MEP consists of exactly two messages, only one of which MUST be a SOAP message:

Chris: still not convinced. is 405 a underlying protocol valid reply?

Noah: should we fix the whole thing or accomodate only marginal uses

<Noah> Noah notes to DaveO that Noah is at least being consistent in not advocating that we fix most of what's suboptimal.

DaveO: I am against doing anything more than the ROR bits

<Noah> NM: Me too. I don't mind fixing very minor "infelicities" when it's easy, but I don't feel like doing this one.

Noah: so is the proposal to avoid clarifying the glossary?

<Noah> +1 to Dave. Ship it.

DaveO: if we need to do the errata, so be it, but it would be good to ship things asap

<Noah> Final proposed SC2 text:

<Noah> The SOAP Request-Response MEP defines a patt

<Noah> The request message MUST contain a SOAP envelope.

<Noah> ******

<Noah> Final proposed SC2 test (trying again)

<Noah> The request message MUST contain a SOAP envelope.

<Noah> ******

<Noah> Final proposed SC2 text (take 3)

<Noah> The SOAP Request-Response MEP defines a pattern for the

<Noah> exchange of a SOAP message acting as a request followed by a message

<Noah> acting as a response.

<Noah> The request message MUST contain a SOAP envelope.

<Noah> The response MAY contain a SOAP envelope, and MUST in all cases include a binding-specific indication that the request has been received.

<Noah> ******

<Noah> In the absence of

<Noah> failure in the underlying protocol, this MEP consists of exactly

<Noah> two messages.

<Noah> ******

<Noah> Final proposed SC2 text (take 4)

<Noah> acting as a response.

<Noah> acting as a response.

<scribe> ACTION: Mike to send to the ML the definitive SC2 text [recorded in http://www.w3.org/2006/04/12-xmlprotocol-minutes.html#action02]

1way SOAP MEP/Binding work item

one editor only, we may have lost critical mass

DaveH: state machines are quite complex, especially in the ROR case as it implicitely describe parallelism, while it would have been better to describe it explicitely. However there is a tradition to describe protocols based on state machine
... so I'm leaning toward using a simple state machine for the one way case

Mike: so you prefer having a state machine (albeit a simple one).So you value 1) consistency 2) rigor

DaveH: yes, it makes things more explicit

DaveO: I do prefer the 1 page-long description of the same thing described by a 3-pages description
... it seems more simple

DaveH: the problem implementing the spec is not reading a nicely written and concise spec, the issue is to understand what the spec means, and state machines (or formal notations) helps.

DaveO: waiting for a formal notation may make you miss the window

DaveH to continue the discussion on the ML.

ADJOURNED

Summary of Action Items

[NEW] ACTION: Mike to send to the ML the definitive SC2 text [recorded in http://www.w3.org/2006/04/12-xmlprotocol-minutes.html#action02]
[NEW] ACTION: Yves to open an issue and send closing email immediately [recorded in http://www.w3.org/2006/04/12-xmlprotocol-minutes.html#action01]