W3C home > Mailing lists > Public > www-html@w3.org > October 2002

Re: OL needs the start attribute

From: (wrong string) šper <christoph.paeper@tu-clausthal.de>
Date: Thu, 17 Oct 2002 18:32:32 +0200
Message-ID: <002c01c275fa$cbd3e2d0$3ef4ae8b@heim4.tuclausthal.de>
To: <www-html@w3.org>

Philip TAYLOR [PC336/H-XP] <P.Taylor@Rhul.Ac.Uk>:
>
> The 3/3a/4 dichotomy can surely be addressed only by
> allowing an individual <LI> not to participate in
> the general numbering scheme; thus one might postulate
> a markup such as the following :
>
> <li ol=none value="3a"> material for section 3a </li>
>
> where 'ol=parent' is otherwise implied, and 'ol="named-ol"'
> is allowed but rarely if ever needed.

But it does belong to the list and is also incremented, just not by an
integer or by a clear algorithm, which is a problem that could be solved
with a user defined list-style-type in CSS3. Btw.: The tra[d|ns]itional
attribute "value" only accepts integers.
Perhaps it's an option to introduce a sub *element* to li (only those in a
ol), which carries fixed ordinal values. That solution would also be
backwards and CSS-off compatible -- nevertheless I dislike it for no
particular reason.

> [I am using here the convention of an unquoted string for
>  a keyword and a quoted string for a variable; not sure if
>  this is in accordance with current or proposed syntax.]


In X[HT]ML *all* attribute values have to be quoted.


For breaks in lists I thought about proposing

  continue="IDREF"

I guess this could also handle multi-document lists, like in

  <ol continue="PreviousPage.html#Part1ofThisList">,

*but* this would require the UA to download and parse one or even more
documents, which maybe weren't of otherwise interest for the user. Thus a
fixed "start" attribute would be more economic but less logic.

Christoph
Received on Thursday, 17 October 2002 12:32:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:53 GMT