Re: WSDL with distint interfaces for return types

On Mon, Nov 21, 2005 at 09:53:37PM -0600, Dan Connolly wrote:
> 
> On Tue, 2005-10-11 at 10:27 -0400, Eric Prud'hommeaux wrote:
> > I've been trying to see what WSDL consumers can do, comparing the
> > current WSDL, with one opertion, to another with two operations, one
> > for each return type. My goal is to sufficiently describe SPARQL so
> > that a tool can set up an appropriate handler for the response.
> > 
> > I'm not done with my homework yet, but wanted folks to see what I was
> > working with.
> 
> Is this proposal still live? I was trying to close off the thread
> with Mark Baker, and I was writing:

it's dead, Jim.

> [[
> We did discuss designs where ASK/SELECT/CONSTRUCT/DESCRIBE
> were split into 2 operations: one for ASK/SELECT where
> the return data is table-shaped, and one for CONSTRUCT/DESCRIBE
> where the return data is graph-shaped. But it didn't get
> a critical mass of support.
> ]]
> 
> but I'm not sure whether folks expected further discussion of
> EricP's proposal. Kendall, are you finished thinking about it?

I had a conversation with LeeF on irc://irc.w3.org:6665/#tp from
2005-10-17T18:16:59Z to 2005-10-17T22:10:24Z . My conclusion is
that tools deal well enough with the current WSDL and that I don't
have sufficient technical arguments to dispute the aesthetic choice
of having fewer operations. This mostly comes from LeeF's summary:

2005-10-17T18:37:09Z <LeeF> that said, although it's a bit logically
                            cleaner, I think the toolkit code would
			    actually be messier

-- 
-eric

office: +81.466.49.1170 W3C, Keio Research Institute at SFC,
                        Shonan Fujisawa Campus, Keio University,
                        5322 Endo, Fujisawa, Kanagawa 252-8520
                        JAPAN
        +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
cell:   +81.90.6533.3882

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.