Re: Lout Backends

[Ludovic Courtès]

Hi !

> [...] As you said, even the question of
> detecting paragraphs at the level of a backend seems monstruously
> cumbersome (or I understood the text backend totally wrong :), or
> impossible.

Yep, that's the point. :) I thought of some ways to detect new paragraphs by
looking at the vpos, but it doesn't seem to be that easy. And the main problem
is that it would really be some kind of a `bad hack'.

As you said, a Lout2html translator could be way more efficient. But the problem
is that it would require almost to rewrite a parser, and it would be also very
specific to a given style-sheet (i.e. doc OR report OR overhead). It would be
very nice if we could use the "@BackEnd @Case" thing to define some
backend-specific behaviour. For HTML, it would also require to bypass all the 
layout
layer, and just have a hi-level communication with the backend (i.e. "the
following text is centered", etc.). But I guess this is not the aim at all of
Lout... ;)

So I think I will just keep trying to reverse-engineer Lout Basser and try to
see if I can manage to do something usable and useful out of it (which again
looks a bit hopeless).

Ludovic.