On Sat, 16 Jul 2005, Michael Thomas wrote:
I agree with a lot of what you say in this post, but I think
you're missing one use case of simple which doesn't have an
dependencies on aspirations: the case were a signer would rather
the signature break -- with even the possibility of discard.
For example, statements(_at_)bigbank(_dot_)com probably does not want anything
monkeyed with their statement, and is willing to tolerate the
risk of manglers. Whether this leads us toward Mark's aspiration
goal I don't know, but I do expect that this sort of hard line
to be common within sections of the deployed base.
That is why intermediate canonicalization method (between your "simple"
and your 'noswp') is good to have. It would not cause loosing of meening
of data if somebody were to see it (i.e. removal of duplicate white spaces
or duplicate empty lines is not likely to be considered esential by some
bank) but would give a lot better possibility of survival with common email
alterations then no canonicalization done at all.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net