W3C home > Mailing lists > Public > uri@w3.org > July 2004

Re: 2396bis - new productions

From: Roy T.Fielding <fielding@gbiv.com>
Date: Mon, 12 Jul 2004 17:43:41 -0700
Cc: "'uri@w3.org'" <uri@w3.org>
To: "Hammond, Tony" <T.Hammond@nature.com>
Message-Id: <B015B6BE-D465-11D8-88C4-000393753936@gbiv.com>

> Well, I'm confused as to what some of these new productions in the -05 
> draft
> (of 2396bis) mean. (Thought everything was more or less hunky dory up 
> til
> -04, apart from allowing dot segment normalizations to be applied 
> across the
> board on all URI forms - both relative and absolute. Though did have 
> some
> general sympathies with this in terms of arriving at a greater level of
> uniformity for the URI spec.)
>
> But now I'm totally lost. What does
>
> 	segment-nz
> or
> 	segment-nzc
>
> mean? ('nz' anyone? - 'nzc'?)

segment-nz means non-zero-length segment.  That should be pretty clear
when provided the ABNF rules in context.  I'll change the segment-nzc
to segment-nz-nc, and path-abs to path-absolute.

    segment       = *pchar
    segment-nz    = 1*pchar
    segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
                  ; non-zero-length segment without any colon ":"

    pchar         = unreserved / pct-encoded / sub-delims / ":" / "@"

However, I don't want people to figure them out by looking
at the terminal names -- I want them to implement the grammar itself,
which is better defined by reading the ABNF rule than by guessing
on the basis of the names used; the names are short because I am
space-constrained to 68 columns.

....Roy
Received on Tuesday, 13 July 2004 16:38:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:34 GMT