> > david is right: to make assertions about the way the user interface
> > "will be", you need to know whether your proposal is actually
> > practicable. if any part of your proposal is not practicable,
> > there's a danger that everything you suggest will be ignored,
> > however good _some_ parts of it are.
> > the fact is, that many people complain about the restrictions that
> > tex places on our programming, but no-one is willing to throw out
> > the basis of the "programming model" of tex -- rebuilding tex from
> > scratch is just too much of a job.
> > :-(
> In the thread about the user interface in de.comp.text.tex Frank
> Mittelbach wrote several times that LaTeX3 is able to separate the
> different levels of the application. In particular he spoke about a
> completely independent user interface layer that enables LaTeX3 to
> support several different interfaces.
> In fact, Frank states that it would not be much effort to change the
> existing LaTeX3 code to adapt it to an improved user interface.
> I don't understand how that fits together with your statements that
> one would have to change TeX fundamentally in order to be able to have
> another interface.
i think you misunderstand.
frank is not talking of "user interfaces that can't be implemented in
tex-as-it-stands" -- no-one would expect that he would, surely? the
problem here is that _you_ don't understand what we can and cannot do
with tex. ambitions to regularise the use of spaces fall into that
category, i'm afraid: one simply can not get "sensible" space-
character processing right.
latex, dating back at least to 2.09 (the earliest version i used), has
some code for ignoring spaces where tex wouldn't, but the user might
expect it to. what you're asking is for latex to reconstruct spaces
that tex has read and thrown on the floor: the best one can hope is
for heuristics about "the sort of way people write their text" (the
xspace package does this, and it's not infallible).