Re: [Texmacs-dev] Multiple column support

[Nix N. Nix]

On Wed, 2003-04-23 at 12:19, Joris van der Hoeven wrote:
> > On Wed, 2003-04-23 at 11:18, Joris van der Hoeven wrote:
> > > I don't agree, because the distinction about a figure/table and
> > > plain text is quite hard to formalize. For instance, the graphic
> > > might be a tree instead of a postscript image, or a colored table.
> >
> > OK, then how about a single object as the body of the figure/table ?  I
> > don't think a paragraph of text qualifies as either a graphic or a
> > table.
> 
> Why not? What exactly is the problem with having a body
> consisting of general text?

I don't have a problem with this per se, however, it makes the width
constraint difficult to implement.  No, this is not reasoning the wrong
way around, because of the followings:

- Figures containing text are rare AFAIK
- Figures are called "figures" because they are meant to contain 
  non-textual information

Consequently, I believe it resonable in this case /and in this case
alone/ to restrict the use of figures in order to better be able to
translate them.

Conversely, if you argue that it's OK for figures to contain arbitrary
paragraphs of text, then you must implement the followings:

- automatic wrapping for the text inside the figure when the figure
width >= line width.

- Next line (i.e., \\) doesn't work inside figures.

I'm not trying to be antagonistic.  I simply believe that having a
single, predictable object inside a figure is a reasonable restriction. 
Please realize that I'm not that adamant about this.  If you don't want
to implement this restriction, that's fine, too.  I will continue to try
to find a way to get LaTeX to tell me the width of a certain chunk of
code, without having to produce that code twice.