Re: ERROR(?) in content declarations for character tags

Murray Maloney (murray@sco.COM)
Fri, 23 Sep 94 12:34:24 EDT

>
> > Please see http://www.acl.lanl.gov/HTML_WG/html-wg-94q3.messages/147.html
> > for my argument in favor of nested emphasis. I think it was widely accepted.
>
> I accept the reason for most of the character tags (i.e. the 'emphasis'
> based ones), but I think it is hard to use the same argument on some of
> the tags. For example, the following tags still, to me, do not make
> sense to allow nesting:
>
> CODE, SAMP, KBD, KEY

First, let me say that I am not entirely convinced by the argument
that I am about to make, but neither am I convinced that these
four elements should be treated specially.

<CODE>:
========

This one is fairly easy to imagine. An author wants to show
a bit of shell script that contains an awk fragment:

<CODE> cat foo | sort -u | awk <CODE>'{ print $1 }' </CODE> </CODE>

The browser could format the nested <CODE> the same as the outer one,
or could chose to add emphasis to the traditional monospace face.

Yes, this is a bit contrived, but since I am the technical publications
architect and strategist for a UNIX Development System, I figure that
I have artistic licence.

<SAMP>:
=======

This one is a bit harder to justify. In fact, I can't.
So, maybe nobody else could justify it either and wouldn't
bother to code a document that way.
On the other hand, since <SAMP> is a "sequence of literal characters",
maybe it should not be allowed to contain anything except #PCDATA.

<KBD>:
======

OK, I can't justify this one either.
So, maybe nobody else could justify it either and wouldn't
bother to code a document that way.

<KEY>:
======

I would imagine that nested <KEY> would be used to code tuplets,
like <Ctrl>D as <KEY><KEY>&lt;Ctrl&gt;</KEY>D</KEY> or <Alt><Ctrl><F1>
as <KEY><KEY>&lt;Alt&gt;</KEY><KEY>&lt;Ctrl&gt;</KEY>F1</KEY>.
A browser could format the nested <KEY>s in a tighter sans-serif face.
>
> Does this make sense:
>
> <code>This <code>is <code>an example</code></code></code>
>
> Code inside code, hmmmm. Saying browsers should just ignore it is not
> a decent argument. I'd like to here why the above should be valid.
> I can't see a logical distinction from a CODE element, that is
> contained within a CODE element, and a non-nested CODE element.
>
> Now this seems perfectly fine to me,
>
> <code>Hit <key>Q</key> to quit</code>
>
> since KEY has a logical distinction from CODE. I'm saying that each
> element above should not contain itself.
>
> --ewh
>