Re: Negotiated noncompliance

[Charles Lindsey <[e-mail]>]

	On Wed, 16 Aug 2000 16:41:38 -0700
	Dave Crocker <[e-mail]> said...

> 
> >  If an implementation receives protocol which is not
> >permitted by the specification and the specification does not mandate
> >the response to such illegal protocol, the implementation MAY consider
> >the peer to have negotiated a nonstandard protocol.
> 
> 
> In other words, a receiver is free to handle errors in any way it deems 
> appropriate, unless the specification explicitly defines the handling of 
> this particular effort.

That indeed seems a messy way to handle it. We know some implementations
are going to do this sort of thing, whether we like it or not.

In Netnews, we have the concept of a "cooperating subnet" who agree
amongst themselves to implement "some variant of the defined protocols".
It is a convenient bucket into which to sweep all those odd little
quirks which individual sites may adopt, and doing so without drawing
attention to them by verbiage visible in the text of the document. That
would be the real damage - drawing attention to possible loopholes will
only encourage people to open them up.

The most I would do is to say that you MUST generate proper stuff, and
you SHOULD NOT accept improper stuff (I agree that people who want to
say MUST NOT accept improper stuff are going too far).
> 
> Besides being on the wrong side of a slippery slope, this sort of language 
> will only serve to make the distinction between "in spec" and "out of spec" 
> more confused than it already is.

Exactly. I oppose the wording suggested, but have no problem with
opening up the discussion in more general terms.

Charles H. Lindsey ---------At Home, doing my own thing------------------------
Email:     [e-mail]  Web:   http://www.cs.man.ac.uk/~chl
Voice/Fax: [phone]      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9     Fingerprint: 73 6D C2 51 93 A0 01 E7  65 E8 64 7E 14 A4 AB A5