[ietf-dkim] 1193 considered harmful

[Eric Allman]

My take is that making gratuitous changes is silly and should be 
avoided.  Personally, I like "|" as a separator better than ":", but 
that's pretty gratuitious and I would oppose such a change on 
principle.

But I'm reminded of the syntax of the sendmail.cf file.  A great many 
years ago I realized that the syntax sucked and I should fix it, but 
I didn't in the name of back compatibility (there were perhaps a few 
hundred servers that would have been affected).  In retrospect that 
was stupid.

If there weren't other required changes then I might feel 
differently.  But the SHA-256 change means that both signers and 
verifiers are going to have to update their software anyway.  This 
will be just part of that update.  No extra work for the installed 
base.

If the proposed change weren't at least some improvement then I might 
feel differently.  But I think it is an improvement.  We can argue 
how much of one, but that's not the point.

If it was impossible to provide a transition period then I might feel 
differently.  But it's not impossible, as has been described.

In short, I'm in favor of this change.

Or perhaps I should have just said "me too".

eric