Posted by Toby Inkster on 05/29/06 00:16

Chris Tomlinson wrote:

> It can't be that hard to do, and I'd be surprised if someone hadn't had a
> crack at it.

Using presentation HTML (e.g. FONT elements, table-based layout) and using
semantic HTML plus CSS are two totally different processes.

A semantically marked up document has more information in it than a
presentationally marked up document. For example:

<h1>Foo</h1>

tells us that "Foo" is a heading for the content below it. "Foo" will
probably continue to be the general topic of all other text until another
H1 element is encountered. Any H2, H3, etc elements further down the page
are probably sub-topics of "Foo".

On the other hand:

<font size=6><b>Foo</b></font><br><br>

tells us far less. It just says that the author wanted the browser to
display "Foo" in some big, bold text; but with no explanation of why this
was desired.

So going from semantic HTML to presentational HTML means that you're
losing information. Conversely, to go from presentational HTML to semantic
HTML means you have to add in some information -- e.g. specify in the
document which parts are headings, which parts are paragraphs, etc. Adding
in that information requires human input; which might be why there are no
tools that can simply transform presentation markup to semantic markup.

Going the other way -- converting semantic markup with CSS into
presentational markup -- might be possible to automate with some tool like
what you describe, but it's not clear why anyone would want to do so.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

[Back to original message]