At 01:50 PM 8/26/2001 +0100, Christopher Tipper wrote:

  >> with some notion of z-coordinates to allow layers to be superimposed
  >> using some notion of priority. Such a concept would not appear to
  >> be possible in the existing TeX architectural model, and is a
  >> significant barrier, in my view to commercial acceptance.

actually, it *can* be done in context sincethere is a layering model, and
actually one can do pretty sophisticated things with it -)

 >Page 1: The line passes through the central float. No doubt some solution
 >could be found whereby a blank white graphic is used as a background to
 >the float, but this is useless if the size of the text box is not known in
 >advance, and must be recalculated for every

This is indeed a good solution, just put a background there, how is the
text typeset now? Put it in a \framed and there you got your background.

 >instance. This is just not sensible, and points out the need for a more
 >general float mechanism. btw I have read Frank Mittelbach's work on
 >xo-pfloat for LaTeX 3, and forgive my obtuseness, but what this guy calls
 >'floats' are just too limited to be useful: for instance I saw no evidence
 >of floats that can be arbitrarily positioned on the page in a
 >deterministic manner.

the new multi column otr will give you some of the control that you want,
including spread graphics and so

 >Page 2: This issue arose during compilation of XML and involves a float in
 >the ConTeXt sense of the word i.e. a graphic. The problem is that the text
 >does not flow accurately past the graphic on the page, and this sort of
 >result would be extremely embarrassing if it were to be published. There
 >is nothing I as a publisher can do to solve this issue, which looks like
 >the result of a flawed algorithm.

\setuptolerance[verytolerant,stretch]

is your friend here.

 >Page 3 and 4: This is more of a question, and may not be an issue at all,
 >but would obviously benefit from a streamlined float handling interface.
 >Page 3 is the XSLT output for the web-page of this article (printed from
 >IE using Acrobat Distiller). Page 4 is the limited rendition I have with
 >the Context from the same XML source. As you can see I have a photograph
 >which has a rather long caption attached. Postioning the caption is
 >trivial in HTML, and I have written XSL templates to transform:

 >   {\placefigure[\XMLpar{image}{align}{}]%
 >     {none}%

 >How may one deal with the <caption> tags in a manner similar to the web
 >layout on page 3?

just put the caption text instead of {none}

or use \startcombination[1*1] {graphic} {text} \stopcombination

this kind of things are doable.

So, live is  not that hard, all you need is a bit of faith: see attachment
for an example of layering; you can imagine that there can be multiple
layers, you can also moves things around on layers, repeat them, etc etc;
[it's a nice mental exercise to figure out hwo this is done.] Now, given
that you can also hook in metapost code, imagine what you can do with
graphics spanning layers.

The issue is not so much if/how it can be done [since much more can be done
that people are made to believe, but to start thinking in how to apply
these techniques. I'm still in the proces of learning.

Hans