Posted by Alan J. Flavell on 01/31/06 23:52

On Tue, 31 Jan 2006, John Salerno wrote:

> Hi everyone. I have a couple of (probably simple) questions about
> how to format some text in a paragraph.

Hint: it's poor usenet strategy to ask two unrelated questions
together. You just get one contorted thread. And the original
subject line is useless too, to anyone who's looking to see if their
question has been discussed before. I've taken the liberty of
addressing one question, and setting the Subject line accordingly.

> First up is the easier one to explain: If I have scientific words
> that I want italicized, is that best done with <em> or a <span>? (My
> guess is span).

Just what do you mean by "scientific words"?

If it's species, for example, then by widely agreed convention they
are presented in italics; but mention of them isn't in itself
"emphasis", so <em> markup would be perverse.

Speaking for myself, in that case I'd go for <i class="species">

Then you get (as far as the client agent can manage it) the widely
agreed typographic convention, even when CSS is disabled; you don't
mislead anyone into believing that it's <em>phasized; and you've got a
class name which, if you wish, could be used for hanging some CSS
onto, or indeed for other purposes such as counting the number of
references to species in a body of work (provided you stick to your
own conventions, of course).

As far as HTML and CSS are concerned, the choice of a class name is of
course arbitrary: class="species" means nothing to them, it might just
as well be class="k42" for all that they care. The choice is purely
for human convenience.

In their appropriate places, don't overlook the other semantic markups
such as <var>, <cite>, etc., some of which happen to default to italic
presentation; but one should never choose a semantic markup just
because one happens to like the default appearance in one's own
browser - use them honestly.

Alternatively, of course, you could define your own rich markup
language, or use one of the existing rich markup languages, for your
internal working, with purpose-designed tags such as <species>, and
then down-convert those into regular HTML, as above, for serving out
to the web. XML-based processing is said to be good for that (I'm not
really into it myself, admitted).

[Back to original message]