2. PEM is too damned late as it is, and proposing "enhancements"
before we have a working version is almost sure to be
counterproductive.
4. Conclusion: at least with respect to the basic algorithm, IT AIN'T
BROKE. DON'T FIX IT! Instead, get it out there and in worldwide use
before the opportunity escapes us.
As one of the folks who proposed to do triple DES for RIPEM, I would second
these two comments. My personal desire was to use RIPEM as a testbed for
triple DES with no intention of having PEM use it. With RIPEM we could
then see if anyone cared to use it. We could play with it without getting
the international standards effort involved.
I would personally prefer to return to that state of affairs -- but I
suppose this has gone too far in pem-dev as it is.
So -- for now, I just second these thoughts -- to leave PEM alone and get
it shipped. Maybe a year after general release, we could reopen this
issue.
Of course, I would still like to experiment with RIPEM.
- Carl