Re: non-outcry over eBay SOAP interace

On Feb 14, 2004, at 8:20 PM, David Orchard wrote:

> I must admit I'm somewhat disappointed in the decision that the wsdl 
> wg took
> wrt this issue.
>
> I would have thought it would be normative and ship in the mainline 
> spec.
> But I guess an extension, even if it won't have any testing or 
> conformance
> work done on it, is better than nothing.

There's nothing stopping any group member from adding tests to the 
notes. If, over the course of developing the note, it became clear that 
there was a solid solution with a community of practice behind it, the 
WG could inline it in one of the other documents, or pursue 
recommendation status for it in a more near term time fram.

> I strongly disagree that there isn't experience in the web 
> architecture wrt
> "safe" operations.

I so didn't say that there wasn't. Indeed, check this quote:

"""It's not like there is, to my knowledge, *any* existing practice for
this sort of description in WSDL. """

And then I reported on some analogous work in SWS for "safe" like 
operations, and using such annotations.

>  In fact, I'd argue that the web architecture is
> completely dependent upon GET being "safe".

As would most web architecture aficiandos ;) Caching, spidering, even 
normal human browsing are dependent on the preponderance of safe 
operations on the web.

>  Because it's safe, it can be
> the default method used to retrieve a representation given a URI.  So 
> we
> have the entire web to show as experience wrt safe operations.  People 
> want
> the same facilities that GET offers but in a more protocol neutral 
> manner.

The advocates of adding safety annotations  were not themselves users 
of them, or people with a user base for them. Given that lack of 
experience, even people sympathetic to the idea were hesitent to put it 
on the normative or critical path.

Is it really so disappointing for the WG to issue a note with a 
recommendation to standardize it? I think standardized WSDL 2.0 
extensions are a key factor in its success. If such annotations really 
are so useful and desired, then there really isn't a stong *need* to 
put them in the core spec at this time. The working group, as I said, 
is ready to reconsider if such a strong need is evident and we can at 
least pull together some stakeholders to participate.

Cheers,
Bijan Parsia.

Received on Sunday, 15 February 2004 07:12:24 UTC