RE: SOAP & REST

Hi Mark,

> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org]
> Sent: 07 February 2002 15:24
> To: skw@hplb.hpl.hp.com
> Cc: xml-dist-app@w3.org
> Subject: Re: SOAP & REST
> 
> 
> Hi Stuart,
> 
> > > The difference is that in the former form, a user doesn't know whether
> >                                                ^^^^person,  browser or
program?
> 
> All of the above.  The URI is opaque.  In the absence of other
> information (i.e. before invoking GET), given only the URI, 
> nobody knows what it does.
> 
> > > the operation being performed is multiplication or division.
> > 
> > Hmmm... if I were using it to calculate my taxes that would seem like a
bad
> > thing to me.
> 
> But you wouldn't use just any URI that you found on a napkin to do
> that job.  You'd use one from some entity you trusted, that when you
> invoked GET on it, described to your satisfaction that this service
> was a multiplier.  Machines can work exactly the same way, though
> obviously they'd need a machine processable assertion of the fact that
> the service was a multiplier, returned on the GET.

Feels like you just contradicted your original premise:

<quote>
The difference is that in the former form, a user doesn't know whether
the operation being performed is multiplication or division.
</quote>

> > > In the latter form, the client is required to specify the desired
> > > operation.
> > 
> > I find this a little perplexing... on one level these are just large
> > 'opague' strings.
> 
> In the former URI, the string "multiply" is opaque to the client.  In the
> latter URI, it is not (see below), modulo the hidden form field kludge I
> mentioned, where the server specifies it as a sort of "callback method" -
> but the *server* specifies it, not the client.
> 
> > You seem to be suggesting that in one case a "client" needs to "know"
what
> > arithmetic operation is "being performed" and in the other case it
doesn't.
> 
> No, not at all.  Obviously, in both cases you need to know that
> multiplication is occuring.  It's *how* this is known that is
> accomplished differently in hypertext and RPC.
> 
> > What I don't understand here is how it can be argued that the sequence
of
> > characters 'm', 'u', 'l', 't', 'i', 'p','l' and 'e'  has any more or
less
> > significance to the client depending on whether it is preceded by a '/'
and
> > followed by a '?' than when it is preceded by the character sequence
> > '/','?','o','p','e','r','a','t','i','o','n', '=' and '"' and followed by
'"'
> > and '+'. 
> 
> See the Opacity axiom, specifically this sentence;
> 
>   "Query strings are clearly not opaque to the client."

I'll think on that a while... 

> MB
> -- 
> Mark Baker, Chief Science Officer, Planetfred, Inc.
> Ottawa, Ontario, CANADA.      mbaker@planetfred.com
> http://www.markbaker.ca   http://www.planetfred.com
> 

Cheers,

Stuart

Received on Thursday, 7 February 2002 11:56:44 UTC