binding.jms: operationSelector, @nativeOperation/@selectedOperation

[Scott Kurz <sc...@gmail.com>]

Did we get this wrong again (my fault)?

In reading the OASIS spec again, I think we're supposed to:
1)  take the service-side operationSelector's selected operationName,
2) look for a @selectedOperation (OASIS) / @nativeOperation (OSOA) in
an <operationProperties> matching the result of 1)
3) determine the value of @name in the same <operationProperties> as in 2)
4) invoke the operation pointed to by @name in 3)

We have it going the other way around... from @name to @nativeOperation.

I don't know why this has taken me so long to grasp... I will say the
new name, @selectedOperation is better.

Scott

Re: binding.jms: operationSelector, @nativeOperation/@selectedOperation

[Scott Kurz <sc...@gmail.com>]

Thanks for stating it so clearly Simon.

You're right.   OSOA was not clear and one could read it either way,
and I'm only making my case regarding OSOA by looking at how OASIS has
clarified this.

I guess I'm just thinking given that @nativeOperation is an
unintuitive choice of attribute name to begin with, I hate to confuse
the user even further by doing a 180 as we go from OSOA to OASIS.

Scott

Re: binding.jms: operationSelector, @nativeOperation/@selectedOperation

[Simon Laws <si...@googlemail.com>]

On Fri, May 1, 2009 at 9:05 PM, Scott Kurz <sc...@gmail.com> wrote:
> Did we get this wrong again (my fault)?
>
> In reading the OASIS spec again, I think we're supposed to:
> 1)  take the service-side operationSelector's selected operationName,
> 2) look for a @selectedOperation (OASIS) / @nativeOperation (OSOA) in
> an <operationProperties> matching the result of 1)
> 3) determine the value of @name in the same <operationProperties> as in 2)
> 4) invoke the operation pointed to by @name in 3)
>
> We have it going the other way around... from @name to @nativeOperation.
>
> I don't know why this has taken me so long to grasp... I will say the
> new name, @selectedOperation is better.
>
> Scott
>

Scott not sure it's your fault I could be persuaded either way. I
think what you are saying is that the OASIS spec is clear in that

OperationSelector -> operationProperties/@selectedOperation ->
operationProperties/@name -> service operation

and so we should interpret OSOA in a similar way where
@selectedOperation is @nativeOperation

I don't think it's clear that OSOA means this but I agree it has a
kind of consistency about it if we make it do that. The use of the
word "native" does make this order slight unintuitive though in the
OSOA case IMO.

Simon