[ietf-dkim] 1193 considered harmful

[Michael Thomas]

Hallam-Baker, Phillip wrote:
>>[mailto:ietf-dkim-bounces at mipassoc.org] On Behalf Of Michael Thomas
> 
> 
>>Define "needed". If the standard of "needed" is "required to 
>>make the protocol viable", then this is not "needed". If 
>>"needed" means "anything we feel like changing, we can 
>>change", then the words in the charter are meaningless.
> 
> 
> Needed means that the value of making the change is justified by the cost of
> making the change.
> 
> In this case the cost of the change now is much less than the cost will be
> in the future. This particular change was proposed multiple times during the
> development of DKIM (I was the proposer). Each time the pushback was the
> cost of making the change.

Really? I don't remember that. In fact, there was no backward
compatibility issues at that time because the DKIM-Signature
header/hashing was different than the DK signature/hashing. As
I remember it, the chosen hash generation was different than both
IIM and DK. And here we are two years later, um, rehashing the
same decision.

		Mike