I was just observing most of this discussion, but I am wondering how much of an issue the whitespace issue will cause when generating signatures. Obviously, different signatures will result when signing an element that has white space and one that does not. Does the canonicalization fix this? Brett.
----- Original Message ----- From: Walter Underwood <[EMAIL PROTECTED]> Date: Friday, August 5, 2005 4:37 pm Subject: Re: spec bug: can we fix for draft-11? > > --On August 4, 2005 9:31:55 AM -0700 Tim Bray <[EMAIL PROTECTED]> > wrote:> > > So for now, I'm -1 on an weakening or removing "The element's > content MUST be > > an IRI" or analogous text in any other section. I'll stop > shouting if I'm in > > a small minority here. -Tim > > Wow, this string has made my "away on vacation" mailbox fatter. > > I strongly favor making white space around IRIs illegal in Atom, > whetherthey are an ID or somewhere else. Same for dates. > > This follows the robustness principle, where we are conservative > in what > we generate. Atom processors are free to be liberal in what they > accept,so they can strip whitespace. Or not, I don't care. > > Note that a feed with whitespace around an IRI can never be aggregated > into another feed, because a) the ID IRI cannot be changed, and b) > the new > feed cannot cannot contain whitespace. > > Making every single processor strip whitespace smells too much like > the HTML tag soup processors that we all have to maintain. Yuk. > > wunder > -- > Walter Underwood > Principal Architect, Verity > >
