ietf-openproxy
[Top] [All Lists]

Re: one more suggestion

2001-08-28 08:03:47

Andre Beck wrote:

Good suggestion. I guess I originally wanted to keep the properties like
"request-line" protocol-independent, but this doesn't really allow for
very fine-grained rule conditions as you pointed out. So it probably
makes more sense to have a set of very protocol-specific properties for
each supported protocol (in addition to any protocol-independent
properties). I will try to reorganize the draft a little bit to reflect
this better in the next version.

We just recently discussed to better structure the draft and to
separate protocol specific elements from the generic elements. The
idea is that IRML defines a generic, protocol-independent "framework",
but allows addition of protocol-specific elements/profiles (e.g. HTTP
header fields for rule matching). We will probably separate out
protocol-specific elements/profiles into an appendix section of the
draft. In this case, Lily's suggestions can nicely be integrated and
would probably simplify rule handling.

-Markus

<Prev in Thread] Current Thread [Next in Thread>