Re: Real layout

Todd Fahrner writes:

 > > The idea that good layout can be distilled to a few rules plus some
 > > declarative annotations intrigues me. As the window size is increased,
 > > the rules adjust the spacing accordingly, but this is interrupted by
 > > phase changes (ice into water into steam) which cause more dramatic
 > > effects, e.g. switching from a single column to a multiple column layout.
 > A major question on my mind is whether the emerging script/object models
 > will deliver this kind of functionality, building on a very simple CSS, or
 > whether CSS itself can (or should) become self-sufficient, allowing for
 > interdependencies like leading being a function of line length, font size,
 > and the ratio of x-height to em of the face.

The percentage unit is CSS can be used to make such relationship, but
it's quite limited: it can only refer to one predefined property. For
'line-height', the predefined property is 'font-size'. This, I think,
is the most natural choice if you have to choose one, but it should
also be possible to take other properties into account. How about:

  P { line-height: max(1.2 * font-size, width * 0.1) }

This would still be within the declarative domain -- I'd like to
remain there, but agree with Dave Raggett that the phase shifts may be
troublesome. In the multicolumn proposal [Shoot, can't find the
referenc now] one can specify the columns either as a width
measurement or as a number that indicates the wanted number of

  P { colums: 10em }
  P { colums: 5 }

When the window is enlarged, the first example will add new
columns. This is a simple example of how phase shifts can be handled.

 > Harumph! Imagine that you're flying over a remote wilderness area in a
 > white gyrocopter, and below you see "S O S" spelled out in twisted metallic
 > debris on a hillside. As you approach you see squallor, corpses,  and the
 > smoke from a DTD rising feebly from a firepit beside a lean-to made of
 > single-pixel GIFs. A few insane survivors brandish their ascii daggers -
 > markup cannibals! And you remark "ghastly! so inelegant! - wings used as
 > letters. And they're using tables for nontabular information
 > representation." So you fly off to seek funding to start an anthropological
 > expedition.... That's what I think of when I hear about how inelegant the
 > current state of web design is. No arguments here, sir! No offense
 > intended, either - I couldn't resist.

Wow! I guess we're on the rescue team..

 > > This seems a very fertile territory for research projects.
 > I'd much prefer a public beta myself!

I'd like both, in that order. A little thinking before the beta
doesn't hurt. The history of the web has too many examples of
premature betas setting precedence.



H      k   o   n      W   i   u   m       L   i   e   W o r l d   Wide  W e b  Consortium
inria # FRANCE

Received on Monday, 27 January 1997 18:11:28 UTC