[ietf-dkim] 1193 considered harmful

[Arvel Hathcock]

 > Six months down the road we have the same discussion about some other
 > compatibility issue and someone suggests that an l=-1

Ah, understood.  That would make for a really ugly final specification 
if it were to occur.

OK, if we don't want to go down that route, I think we can still handle 
this problem because I'm convinced that our verifiers can be taught to 
detect when the new hashing mechanism is in use and act accordingly. 
This requires a verifier update but that's already unavoidable.

-- 
Arvel