RE: HTTP features w/ low 'implemented' and 'tested'

• From: Larry Masinter <masinter@parc.xerox.com>
• Date: Sun, 29 Mar 1998 01:49:26 PST
• To: "Roy T. Fielding" <fielding@kiwi.ics.uci.edu>, "Life is hard... and then you die." <Ronald.Tschalaer@psi.ch>
• Cc: HTTP-WG@cuckoo.hpl.hp.com
• Message-Id: <000201bd5af7$f5d2d960$e3d3000d@bronze-208.parc.xerox.com>

>Concerning the low implemented list:
>
>    H 10.1.2 101 Switching Protocols
>    H 14.42  Upgrade
>
>I'm not sure what there is to implement here... If any server does
>implement these, then what do they accept in the Upgrade header?

> Personally, I find it incredibly shortsighted for the IETF not to
> include a special exception for optional features intended to support
> future protocol extensions and/or replacement.  I'd hate to have to
> exclude good engineering practice just to call it a draft standard.

Surely we can find two different people to try to implement
Upgrade and Switching Protocols and test their implementations
against each other.

If I write a client that attempts: 

 Upgrade: HTCPCP/0.0104

can we get someone to create a server that returns 101 Switching Protocols
instead of 419?

Larry

Received on Sunday, 29 March 1998 01:51:46 UTC