Re: changing i to em

[email protected]
Tue, 18 Jul 1995 16:28:44 -0400


[email protected] wrote:
> I don't know if SGML supports the concept, but couldn't certain throwback
> tags like <i> and <b> be interpreted in the new standard as aliases for
> <em> and <strong>, with the goal of deprecating them in a future standard?

No, because they don't mean the same thing.
EM means something is emphasised, as in
I'm <em>very</em> sorry to hear of your unfortunate loss.

But italic might also be used for the title of a sculpture or work of art,
or for an alien term or abbreviation such as <i>op. cit.</i> or <i>passim</i>.

In SGML DTDs, I sometimes do something like
<kw what="place">London</kw>
if there are lots of reasons why something might be in italic, and it isn't
practical or appropriate to write a DTD to restrict them. Then if there are
lots of <kw what="alien term"> elements, I can add an <alienterm> element to
the DTD later without creating a legacy problem.

The attribute can either be CDATA, or (as I prefer), a list like
What (person|place|alien|abbreviation|title|other) #REQUIRED
If.Other CDATA #IMPLIED
where If.Other is used only when What=other. There's no way to enforce the
only-if-another-attribute-is-set-to-other behaviour in a DTD, although an
editing application could do so.

However, what is suitable for scholarly markup is not necessary appropriate
for the list of kooool sights wot me an' my mate saw wiv netscape, eh?

Lee