- From: David W. Morris <dwm@xpasc.com>
- Date: Wed, 9 Jul 1997 22:53:34 -0700 (PDT)
- To: Donald Neal <d.neal@waikato.ac.nz>
- Cc: http-wg@cuckoo.hpl.hp.com
On Thu, 10 Jul 1997, Donald Neal wrote: > At 08:26 PM 9/07/97 +0200, Koen Holtman wrote: > [...] > >I have no problems with people putting creative protocol extensions > >which violate a MUST in HTTP/1.1 in their proxies. I would only have > >problems if people would go around distributing, or making available, > >these proxies as being fully HTTP/1.1 conformant, without telling > >anybody about the extra `special stuff'. > > > >We need to draw a firm line between `plain' and `extented', else there > >will be all kinds of trouble when cascaded proxy networks which span > >multiple organisations are going to be built. > > Such networks already exist, of course. Try asking the same question in > another way. If I sign an agreement with another cache operator in which we > state that our proxy caches are HTTP/1.1 compliant, does that contract > allow either of us to alter our proxy caches in such a way that one or more > of the MUSTs are violated? Thats fine ... but this thread started with Dave Kristol's objection to the addition of a MUST which objection I supported. So I wish to avoid unnecessary MUSTs and your contract will hold. Dave Morris
Received on Wednesday, 9 July 1997 22:55:21 UTC