WP>[...]. emacs-w3 will gracefully switch
WP>styles wherever it encounters the <style></style> container, but I don't
WP>know about arena.
AH>I think the issues, especially for the slow connection users go
AH>beyond this. If we have to transmit a couple K of style info
AH>throughout the document when we really just want to have a
AH>formatting tag or two (And once again I am talking about
AH>"lightweight" usages rather than "corporate publishing"), it is an
AH>incredible waste of time for slow speed connections.
Alex recommends a tag, <C>, with attributes FONT, FONTCODE, TYPE,
COLOR, and BGCOLOR. Font specifies a named font ("times",
NCSchoolBook", etc..), FONTCODE specifies a type of font ("serif",
"sanserif"), and COLOR specifies the text color. (I'm including this
because I had trouble getting through to his site.)
The <style> tag as implemented in w3 mode would allow the same sorts
of changes by giving the changes as deltas to the current style.
Here's (probably) an example of changing the text to red helvetica in
each of the style syntaxes. Note that they use a comparable amount
of bandwidth.
<C FONT="helvetica" FONTCODE="sanserif" COLOR="#ff0000">
<STYLE>
font.family = hevetica
font.color = #FF0000
</STYLE>
Personally, I prefer the <STYLE> method because it is more expandable
than the <C> tag -- it doesn't require additional attributes to a tag
to permit new functionality -- all the functionality of the proposed
stylesheet mechanism is already there.
However, I agree with Joe's observation in
<9505022220.AA12225@trystero.art.com>, and think that using the
<STYLE> tag like this should be allowed, but deprecated. The
preferred mechanism for doing this should be using the class tag as
shown at
<URL:http://www.w3.org/hypertext/WWW/Style/css/draft.html#class> and
leaving the <STYLE> tag in the head.
This has the advantage of making it easier to chop the document into
fragments using <DIV> or other mechanisms without risking the loss of
style information.
As a separate issue, one question I would like to ask about the
specification is if the syntax
*.class_name: something.or.other
is allowed. I've looked through archives and various documents
looking for an answer to this. It would seem to make sense and
to be generally useful.
--Corp
Speaking only for myself