Re: Positioning HTML Elements with Cascading Style Sheets

At 5:38 PM -0800 2/1/97, Taylor wrote:

> Using gifs to to position things doesn't work all that well is the least
> degradable way to do things, and allows for no considderation of what I'm
> using to view the document.  Tables are such a complex layout device that
> only the human mind can really use them to lay out a page in the most
> effective manner, and they are also a hack that does not degade well if
> are using them for presntational purposes.

I could not agree more with you about the problems of the current
gif-and-table design regime. What did Churchill say about democracy - that
it was the worst conceivable system except for all the available

> Positioning allows tools that write out good html, and allow the author to
> decide upon what the page looks like.  As for a good degradation strategy,
> that is where client side scripting comes in.

I have a problem with this. Client side scripting is (a) where the browser
war battle lines are currently drawn and (b) where document authors and
graphic designers tend to lack the requisite programming skills. If CSS was
to end the tag wars, client-side scripting will only replace them with more
damaging script wars. Effective scripting, much more so than effective
table structure, is very unlikely to be generated by a robust WYSIWYG
authoring tool. That's why I favor enriching CSS, allowing some values to
be simple functions of others, and a mechanism for allowing alternate
stylesheets to apply dynamically under varying rendering conditions (such
as changing window sizes, speech synthesizers, printers, etc.)

I may be wasting my time, what with Netscape's J-ESS (or JASS or JSSS or
whatever) and Microsoft's Trident headed for public beta. The former of
these technologies especially demonstrates that if you're comfortable with
Javascript, you can bypass CSS altogether. I fear that the programmers who
are bringing us script-based solutions to style (J*SS, CSSOM, even DSSSL)
have been conditioned by their education to deprecate the declarative,
procedural nature of CSS in favor of Turing-complete approaches. Yet if the
popularity of <I> over <EM> is any indication, authors are *more
comfortable* with declarative, procedural systems than more abstract ones,
however powerful.

Todd Fahrner

The printed page transcends space and time. The printed page, the
infinitude of books, must be transcended. THE ELECTRO-LIBRARY.

--El Lissitzky, 1923

Follow-Ups: References: