> What's the difference between "first spec'd binding" as you used those
> words above and "mandatory" as I described?  My point is that a working
> group is going to have to craft a spec that allows two implementations to
> interoperate.  If someone implements something using http, and someone
> else does something different (and the specs allow both) such that they
> don't interoperate, then there is going to be a problem.  The foundation
> that allows interoperability MUST be specified in the charter.

I'm arguing that the foundation for interoperation is the DIX protocol and
not the binding to a specific transport protocol (HTTP). By making a
particular binding "mandatory" for all implementations, you are potentially
limiting future choices. For instance, would I be forced to implement HTTP
in all XMPP and SIP clients and proxies even though XMPP and SIP bindings
may exist? And that begs the question - why HTTP?

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Scott
Hollenbeck
Sent: Friday, January 20, 2006 4:13 AM
To: 'Digital Identity Exchange'
Subject: RE: [dix] draft of proposed charter (#2)

> -----Original Message-----
> From: Suresh Venkatraman [mailto:[EMAIL PROTECTED] 
> Sent: Thursday, January 19, 2006 10:34 PM
> To: 'Digital Identity Exchange'
> Subject: RE: [dix] draft of proposed charter (#2)
> 
> > I think Scott's point is that we mandate that the parties must  
> > implement one particular binding.
> 
> That's the part I was disagreeing with. I was just trying to 
> pointing out
> that the first spec'd binding, HTTP, would allow for DIX to 
> build traction
> and that we did not need to have a "mandatory" binding.

What's the difference between "first spec'd binding" as you used those words
above and "mandatory" as I described?  My point is that a working group is
going to have to craft a spec that allows two implementations to
interoperate.  If someone implements something using http, and someone else
does something different (and the specs allow both) such that they don't
interoperate, then there is going to be a problem.  The foundation that
allows interoperability MUST be specified in the charter.

-Scott-


_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix



_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix

Reply via email to