On July 17, 2005 at 15:27, domainkeys-feedbackbase02(_at_)yahoo(_dot_)com wrote:
Do you want to advocate one of these mechanisms in conjunction with DKIM?
This sounds like (to me): "We already have developed code that does
simple and nowsp and we do not want to take the cost to change it."
Sorry if I am being short.
I think the source of criticisms about the canonicalization algorithms
in the DKIM draft is the apparent lack of consistency with existing
mail standards wrt to this issue.
For example MIME goes to great lengths to address common mutations
(hence, QP and base64 encodings). The nowsp algorithm in DKIM appears
to try to address the issue, but it goes to far; I have already
given my reasons why and an alternate algorithm which I believe is
more appropriate.
RFC-2822 explictly states that header fields names are to be treated
case insensitive and the header fields are to be unfolded before any
further processing is done on the field.
I have yet to see any convincing arguments to go contrary to these
specifications. And if they do exist, they should be explcitly
mentioned in the DKIM specification so adopters clearly understand
why existing practice is not being followed or modified.
--ewh