Re: Further deprecating PGP2

[Adam Back <adam@xxxxxxxxxxxxxxx>]

Hmm I am missing a couple of negatives here for this to make sense,
this is better:

One could hardly consider an implementation bad that encrypted with
IDEA when IDEA is the only defined algorithm for that key (and this is
what I'd read SHOULD NOT to mean; MUST NOT would make the
implementation non-conformant even!)

Adam

On Tue, Mar 11, 2003 at 12:06:01AM +0000, Adam Back wrote:
> 
> I agree.  SHOULD is fine, but if people have strong feelings about it
> MAY is ok also and has similar effect for implementations that care to
> add backwards compatibility.
> 
> One could hardly consider an implementation bad that failed to encrypt
> with IDEA when IDEA is the only defined algorithm for that key (and
> this is what I'd read SHOULD NOT to mean; MUST NOT would make the
> implementationon-conformant even!)
> 
> Adam
> 
> On Mon, Mar 10, 2003 at 06:34:14PM -0500, Derek Atkins wrote:
> > 
> > David Shaw <dshaw@xxxxxxxxxxxxxxx> writes:
> > 
> > > Even better, even word games can become tiresome.
> > > 
> > > What do you advocate with regards to IDEA - SHOULD NOT?  MUST NOT?
> > > something else?
> > 
> > I would NOT advocate MUST NOT.  In fact, I wouldn't even leave it as a
> > foo-NOT.  I'd still leave it in the positive.  I'd still rather keep
> > it as 'SHOULD', but 'MAY' is reasonable.  Why is IDEA so detrimental
> > that you MUST NOT use it?