[ietf-dkim] 1193 considered harmful

[Michael Thomas]

Paul Hoffman wrote:
> At 11:29 PM +0000 3/21/06, Stephen Farrell wrote:
> 
>> This is not a gratuitous change since there are folks who
>> believe it offers sufficient advantage.
> 
> 
> Exactly. People have listed multiple good results from the proposal. So 
> far, the only negative result is that people who have written software 
> to the -00 draft will have to change it.

I guess that's an inconsequential consideration.

But you forget to mention that the combinatorial explosion problem that
results with each incompatible change. And that's doubly true since the
bar is being set so low for any subsequent incompatible change. And let
us not forget that the single most difficult thing to achieve
interoperability is the canonicalization/hashing.

		Mike