Google Groups

Re: Where is LaTeX/TeX headed?

Peter Flynn Apr 8, 2012 7:23 AM
Posted in group: comp.text.tex
On 08/04/12 12:14, Marc van Dongen wrote:
> On Sunday, April 8, 2012 11:26:44 AM UTC+1, Robin Fairbairns wrote:
>> however, i've seen incremental typesetting in latex happening
>> (lightning textures).  i don't _think_ i would find it difficult,
>> despite the occasional jumps in the display.  but without the
>> resources to put the whole "together", i can't say, for sure.
>> which is a pity.
> I don't think fully incremental typesetting in LaTeX is
> realistic---not even in the next decade.

If you mean character-by-character re-typesetting, as Peter D pointed
out, I agree, although perhaps quantum computing might provide us with
instantaneous re-typesetting on that basis :-)

But for the average user I don't think that's the problem. I suspect
that displaying the part-completed paragraph during typing in
line-by-line mode would be perfectly adequate; after the user hits Enter
or whatever to terminate the paragraph, it could asynchronously
re-typeset itself and gently re-display itself, possibly with a fade to
indicate that it is optimising itself. Or that could be suspended at
user discretion, to be implemented later. Subsequent re-editing of the
text would re-initiate the same routine.

It's important to understand here that I am NOT talking about the use of
[La]TeX by the professional compositor, the wizard or guru user, the
deeply-experienced user of many decades standing, or the typographic
expert. I am talking about the average person wanting to create a
document, whether in academic, business, government, NGO, or at home.

> For example, some documents
> need _several_ runs before the output document is finalised. For
> example, consider a document with several multi-page tables. The
> first table may occupy $n$ output pages in the first couple of runs
> and $n+1$ in the next run. This may force the next table to the next
> page, and the same scenario may now apply to the next table. And so
> on.

I think in any synchronous scenario, the algorithm for floats is going
to have to be revisited. In any event, users (of all systems) are
already used to a document re-arranging itself *while it is still being