Re: Issue 169: Propose http method in the operation interface to simplify http binding.

Hi Mark.

* Mark Baker <> [2004-06-29 21:21-0400]
> This proposal seems to me to reopen issue 64.  Correct?  If so, what's
> to become of Hugo's proposal[1]?
>  [1]

Thank you for refreshing my memory, as with all the discussions around
the HTTP binding, I had lost track of this issue.

The issues list states[2]:

  Issue 69 resolution: Adopt Hugo's proposal; open syntax issues 169,

So, I think that at this point the debate is on the syntax at this

My proposal was essentially to specify the Web method with a feature,
specifically the SOAP Web Method Feature. There are two issues with
this approach:
- Some people don't like the idea of reusing the SOAP Web Method
  Feature as there's SOAP in the name and in the URI. I have recently
  reread the feature specification[3], and the fact that it associates
  it with certain SOAP MEPs made me doubt about how generic it was.
- A feature allows this information to be specified potentially at any
  level, and in agreement with Dave and Mark, and unlike Amy, I
  believe that this information makes sense at the interface operation
  level, and Dave is right on by citing Atom as an example (Atom
  should make use of the SOAP Web Method Feature IMO, but that's
  another debate). I don't believe that this is HTTP-specific: SOAP
  supports this concept, and other protocols that we may want to bind
  the operation may also.

Therefore, I am happy with Dave's proposal, along with the defaulting
mechanism he proposes.

I would propose the friendly amendment to tie the webMethod attribute
to the SOAP Web Method Feature, by adding a paragraph in Part 3, in
the SOAP binding, saying that the value of the webMethod attribute
sets the value of the SOAP Web Method Feature property.



Hugo Haas - W3C -

Received on Wednesday, 30 June 2004 05:34:50 UTC