- From: John Cowan <cowan@mercury.ccil.org>
- Date: Wed, 28 May 2003 23:10:48 -0400
- To: "Roy T. Fielding" <fielding@apache.org>
- Cc: uri@w3.org
Roy T. Fielding scripsit:
> >Section 2.4.1 does not state that escapes in the range %00 to %7F are
> >to be understood as US-ASCII encodings, which seems unnecessarily
> >vague.
>
> That is defined by the ABNF specification, with the proviso given in
> section 1.3. Section 2.1 repeats it.
>
> >It should further be specified that the mapping of escapes in the range
> >%80 to %FF is not defined by this RFC.
>
> Why? The RFC doesn't define any such mapping.
>
> >Furthermore, it is not specified
> >that the second and third octets of the escape correspond to the high
> >and low order nybbles of the US-ASCII value and not vice versa.
>
> They are not nibbles -- it is a number in hexadecimal.
These comments refer to escapes within the URI, not to the conventions
of ABNF. As things stand, "a%82c" might mean "abc" on an EBCDIC
system, which seems perverse to me. Still worse, there is nothing to exclude
"a%26c" as equivalent to "abc".
> The algorithm certainly hasn't "prevailed" -- browsers consistently
> ignore that rule and represent the target as base-URI#fragment.
Experiments indicate that you are correct. It is a bit dodgy to
change the standard to conform to the properties of browsers, though.
--
John Cowan jcowan@reutershealth.com www.ccil.org/~cowan www.reutershealth.com
"If he has seen farther than others,
it is because he is standing on a stack of dwarves."
--Mike Champion, describing Tim Berners-Lee (adapted)
Received on Wednesday, 28 May 2003 23:11:14 UTC