[DISC] Re: reference implementation(s)?

[Dave Kormann <davek@research.att.com>]

> I'll endevour to ensure it passes the protocol cops tests.

speaking of which, it'd be useful to have something like this; in
addition to reference implementations, it'd be good to have a
certification rig that abused implementations and certified them as
legal.  i think probably something that did, among other things:
- going through a startup sequence
- sending bogus messages and args, and expecting them dropped
- sending multilines
- receiving multilines
- testing cords, if they're supported
and so on

we'd want this for both server and client implementations.  we'd
probably need to agree on some base protocol, something like maybe an
echo protocol, to do this.
				dk