[rfc-i] ABNF (RFC2234) vs HTTP's augmented BNF syntax (RFC822 + RFC2616)

[Bob Braden]

  *> 
  *> So the summary would be that it's ok to invoke the "implied LWS" rule 
  *> (after all, it doesn't affect the ABNF syntax, just its meaning in the 
  *> spec), but one should stay away from defining extensions (such as 
  *> RFC2616 does to RFC822)?
  *> 

Julian,

Well, I do not mean to say that this is necessarily the last word on
the subject, but at this point, the conservative ("standards-based")
approach you outline does seem to have significant advantages.

Bob