Yes Dale, good summary. We have already discussed this with the XMLDsig folks and they informed us they identified this problem and decided to let the applications deal with it! Maybe something more can be done. I don't see our list of transforms to be an exhaustive list. If someone wanted to add another transform in their implementation, I don't think that would break our spec. If the Dsig group fixes this problem, implementations could use the new canonicalization method they define. Rather than try to resolve this at this eleventh hour, I have inserted a warning in the Multi-Hop section. Regards, David.