- From: Frank Ellermann <nobody@xyzzy.claranet.de>
- Date: Fri, 18 May 2007 00:44:17 +0200
- To: www-html@w3.org
Tina Holmboe wrote: > (a) All presentational elements - and yes, this does include I, B, > FONT, and the way M is defined today - are removed from the HTML > specification. Breaking backwards compatibility for something that's supposed to work, if that's your vision of progress PLEASE don't abuse text/html for the outcome. Some cases about <i>, <b>, and related issues I'm aware of: Some wikis support '' or ''' or ''''' (5) constructs. About a year ago this was intentionally rendered as <i> or <b> or (IIRC) <i><b>, because the Wiki can't guess what the author meant, <em> or <strong> or what else. I can't say if that's what they still do today. They certainly used XHTML 1.0 transitional like everybody else interested to get it right for any browser, strict doesn't help wrt backwards compatibility. Another wiki supports '' or ''' or '''' (4) constructs, rendered as <em class="em2">, <em class="em3">, <em class="em4">, or similar. That gives me no difference at all with my CSS-agnostic browser, therefore I use <strong> directly instead of '''. Probably causing havoc for users of this Wiki when they try to roll their own personal style sheet. The "meaning" of <i> and <b> is perfectly clear: <i> means "somewhat different from the ordinary style, but not much", <b> means "like <i> but never the same as <i>". With a minimal curses model (killing all fonts and funny colours on the fly) of 7 colour pairs <i> and <b> need 3 out of 7. A decree that <i> matches <em> and <b> matches <strong> is fine for this purpose. Add <u> matching whatever <a> does, and <s> matching what <strike> does, deprecating the dangerous <del>, and you're at 5 out of 7. That leaves one for <tt> matching <code>, and the last is up for anything you care to name, <small> or <shipname> or <big>, it doesn't matter. Monospaced text with 7+1 colour pairs isn't excessively limited, it's more than good enough for simple pages, it's also more than good enough for speech browsers trying to render different things differently, what might they do, voice male / female, fast / slow, loud / silent. It's also more than good enough for simple B+W printers, when they try to interpret these styles literally as italics, bold, underlined, etc. Of course it's presentational in some sense. If you want to add stuff like <shipname> you'd have to define how important it is to get effects clearly different from say <em> on _any_ device supporting differences at all. That's what some authors want, clear differences. And it has to work with the oldest version of Lynx you can still start on your box. Frank
Received on Thursday, 17 May 2007 22:51:37 UTC