On Fri, 02 Jan 2004 13:14:54 +0100, Julian Reschke <julian.reschke@xxxxxx> said: > The second statement seems to be misleading. For instance, RFC2396 (URI > syntax) updates previous RFCs that also contained specific URI scheme > descriptions (such as "ftp"). Thus, it doesn't obsolete them. However, > RFC2396 clearly can be used without the documents it's updating. OK. Maybe I'm just caffeine-deficient today, but I'm totally failing to understand how you can use the parts of rfc2396 that update the rfc1808 definitions of relative URI's, if you're not using relative URI's as specified by 1808. Or is your comment addressing the fact that although some sections of 2396 are dependent on 1808 and are left dangling if 1808 isn't included, that *other* sections of 2396 are stand-alone and not dependent on it? (i.e. if your situation doesn't use relative URI's, you can disregard both 1808 and the update to it)? If so, there's likely a little bit of confusion between what the standard actually specifies (a large number of URI classes), and what any given application may actually implement. So while for you *as implementor* large sections of 2396 may be irrelevant, in order to make the *standard* consistent, it must be dependent on 1808 (otherwise, some *other* coder who *does* have to worry about relative URI's will get burnt). Does that clarify things any?
Attachment:
pgp00392.pgp
Description: PGP signature