Ballot on Revised DSA mechanism proposal

[Michael StJohns <msj@nthpermutation.com>]

Point of order I think?

I went back and reviewed the bidding on this proposal - it's current 
5/2/1.  But almost all of the yes votes are "contingent" on fixing a 
bunch of stuff.  My understanding of the process is that typo's and such 
can be fixed, but that substantive changes are pretty much forbidden.

As I read the proposal, (and in fact even a few of the comments with Yes 
votes seem to read this way) there are substantive changes needed not 
just editorial.  I'm particularly nervous about backwards compatibility 
- specifically CKA_SUBPRIME_BITS is now (per the document and per 
Relyea's comment) a required attribute for parameter generation.   It 
wasn't a required attribute for parameter generation for 2.20 or 2.30 
(and I'm not even sure it was a permitted item).  That means that a 
current application that currently works generating DSA parameters is 
going to get an error if it attempts to generate against a 2.40 HSM.

AFAIK we're trying not to do backwards incompatible changes in this go 
around.

I'd really like to ask those who have voted yes or planning on voting 
yes to consider this point specifically as I don't think it can be fixed 
as an editorial change.

Mike