Present
- Glen Daniels
- Henrik Frystyk Nielsen
- Noah Mendelsohn
- Hugo Haas
- Chris Ferris
- Doug Davies
- Mark Nottingham
- Highland Mary Mountain
- Stuart Williams
- David Fallside (scribe, chair)
DF: summarises comments made in XMLP f2f meeting with regard the "SOAP
Underlying Protocol Binding Framework" document,
-- the proposal represents the right direction, as far as it goes, but it
does not go far enough to warrant incorporating into a WD at the present
time
-- the Framework does not provide enough prescriptive statements
-- the Framework is too abstract and does not provide concrete instances or
examples of its usage
The expression of these general sentiments at the f2f were confirmed by
several on the telcon (e.g. DD, CF, HMM) who had voiced concern about the
Framework at the f2f meeting.
NM: noted that sometimes not saying too much in a spec is more important
than saying too much. Also suggested it may be difficult not to
overinterpret a spec.
GD: the Framework should include examples
HFN: the Primer might be a good place to provide examples
DF: what examples, concrete instances, should be incorporated into the
Framework, or elsewhere, that would satisfy the concerns raised at the f2f
and enable us to move forward?
SW: suggests Message Exchange Patterns (MEPs) that are applicable across
multiple bindings
CF: suggests features and properties. Asks whether we intend to describe a
canonical set of features/properties.
NM: reminded us that 3-4 weeks ago the TBTF had backed away from a design
that explictly described a canonical set of features/properties
HFN: perceives we are building a _framework_ of general mechanisms, and
that particular features/properties would be examples rather than
describing a canonical set
GD: we are tackling a complex problem!
NM: suggests we are designing a framework that describes how to define
features/properties, and in fact designing a binding -- the thing that says
how to move bits from A to B -- is actually easy
DD: the distinction between features/properties and bindings suggests to
him that the TBTF should be split into 2 activities
Discussion on what is meant by a "core" feature/property, and whether it is
"important" or "mandatory" in the spec senses of the terms
CF: wants to be sure that within a binding and for any particular binding,
there is sufficient specificity to be useful for implementation
The TBTF decided to schedule another meeting for 24 Sept, and to three
activities:
1. SW to write-up an abstract definition of at least a request/response
MEP, and possibly a one-way MEP.
2. HFN to (re-)write the HTTP binding and cast it in terms of the
Framework.
3. All members of the TBTF to look at how the Framework description might
be clarified.