On 2/11/2015 23:30 pm, Bryan Ford wrote:
2.3 also good, I'm very keen on that. The "bucket expansion" scheme is likely
to signal which tool was used, unless we can convince other packages to do that (pretty
unlikely).
Great. My hope is that if we were to specify the padding/bucket-expansion
mechanism in a separate document in an application-neutral way and with the
relevant theory spelled out, we might eventually be able to convince other
applications to use or migrate to such a scheme too. But that would be a
long-term goal, and whether or not it happens it would have to start somewhere,
and to me OpenPGP seems like a reasonable place for it to start. ;)
Right that's what I was thinking - write a separate draft listing an
algorithm to determine next padded length to aim for.
I'd actually suggest slightly table driven:
512
1024
1500 (ethernet or the old UDP packet length)
4096
16k
64k (jumbo packet)
...
iang
_______________________________________________
openpgp mailing list
openpgp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/openpgp