[xml2rfc] 1.31 formatting nit

elwynd at dial.pipex.com (Elwyn Davies) Wed, 30 August 2006 06:00 UTC

From: "elwynd at dial.pipex.com"
Date: Wed, 30 Aug 2006 06:00:08 +0000
Subject: [xml2rfc] 1.31 formatting nit
In-Reply-To: <44F58133.9060908@gmx.de>
References: <c1468ac50608260546u59e5c80aj2ef5391b2c055de8@mail.gmail.com> <44F059AC.3090908@dial.pipex.com> <44F06CB8.7010508@gmx.de> <200608301203.k7UC3KTE010320@bright.research.att.com> <44F58133.9060908@gmx.de>
Message-ID: <44F58C8D.3090300@dial.pipex.com>
X-Date: Wed Aug 30 06:00:08 2006

Julian Reschke wrote:
> Bill Fenner schrieb:
>>> 2) That being said, people want paragraph breaks in list items. We 
>>> really should fix the rfc2629 DTD in some way to allow this, so 
>>> people don't have to fall back to ugly hacks such as <vspace> (which 
>>> is presentational, not semantical markup).
>> I think this is sensible to brainstorm about, since it's something that
>> has been asked for a lot.  Something like:
>> <!ELEMENT list  (t+|lt+)>
>> <!ELEMENT lt    (t|list)+>
>> <!ATTLIST lt
>>           hangText    %ATEXT;            #IMPLIED>
>> Single-paragraph lists could still use <list><t>... for backwards
>> compatability, new lists would be formatted with <lt>..</lt> wrappers
>> around each item:
>> <list><lt><t/><t/></lt><lt><t/></lt></list>?
>> (A list should only be allowed to have t+ or lt+, not a mixture,
>>  to avoid confusion (always wrap or never wrap))
> Sounds good to me.
> Best regards, Julian
<lt> is a good idea.

The (always wrap or never wrap) requirement is less clear to me.  It is 
related to the confusion (and indignation) of my colleague who could not 
see why <list> should not be a child of <section>..

 From a semantic point of view a list may or may not be closely related 
with the text that precedes or follows it to the extent that it should 
either be in a separate element or embedded in the <t> containing the 
pre/post-amble.  If the list is really a separate semantic entity then 
having to embed it in another (otherwise empty) <t> seems to be the tool 
imposing on the user rather than allowing the user to do what looks natural.
(Think about what you would want if lists had collapse/expand 

Similarly with lists of lists - is the reduction in confusion.more than 
the loss of naturalness?

One view would be that actually everything in a section is a list, and 
<t> is just a special degenerate case.