On Mon, Sep 22, 2003 at 08:51:20AM -0400, Kuba Ober wrote: > > That's the problem. The whole idea of the digestive tract is broken in > > my not so humble opinion. > > Uhm, OK, so you're really looking at reimplementing something that gives you > TeX power yet is more sane to use.
Well, I am not. I am just not calling TeX (the language, not the typesetting as such) 'the best thing since sliced bread'... > A noble goal. I smell LISP if I were to do it cleanly; most probably > LISP with a "saner" syntax so that things are not all just brackets. There are 'bracketless LISP dialects', but that's not the point. Any method cleanly separating 'commands' from 'data' probably will be doing fine. > > [In case you are interested in a puzzle: Set up TeX such that the > > first page of output will have a textwidth of 10cm and every other > > 15cm for any input. One-pass solutions prefered, but mulitple-pass > > is acceptable] > > ROTFL. This is almost like writing TeX from scratch ;-) I don't even > feel like thinking about details of that. It took me enough time to > figure out how to do simpler things :). It would involve resetting the > paragraph box that spans page1/page2 with artificially ragged right > edge. Belch :) But it's not an outstanding request to a typesetting program to format a specified page a bit differently, is it? > I think that TeX (and Qt's richtext engine as it is) becomes very hard > to use due to its bottom-up approach to sizing boxes. Namely that > mostly a box's size is either known a-priori, or it will be known as > soon as horizontal mode is done with. This is something that's bound > by how electromechanical typesetting machines might have worked, but > it's definitely not the most natural way to describe the layout. This > makes e.g. maintaining a certain aspect ratio (width/height) of a > paragraph box a nightmare to implement. And so on. The problem here is that you can't 'undo' paragraph breaking. Why? I don't know. There is no theoretical technical limitation except that TeX simply can't handle it. > > Before I did that, I would try to have a 'sanitized' TeX (i.e. try > > to keep much of the sytax except for the quirks but replace the > > digestive process with something suitable for machines and human > > consumption...) > > I'm thinking whether it would be possible to have a 100% > TeX-compatible thing that would still offer new (and hopefully saner) > syntax and better introspection. Multipass-based "introspection" is a > misnomer, I admit. I think so. But even so, I would not try it and rather settle on something tex2lyx currently does: handle 'sane constructs' which make up the majority of constructs used. > I was hoping that maybe you had ideas how to keep the digestive tract > intact yet make significant improvements. The idea of (TeX's variety of) macro expansion, 'mouth' and 'stomach' is more than my digestive tract can stand for a longer period... Andre'
