From: Ryszard Kubiak <rysiek@bi...> - 2005-10-28 09:44:19
I would like to anounce that I've implemented
a solution for the paragraph problem. Although
it's very pre-mature at the moment it has started
to produce reasonable results. Please compare
the test tex source and the generated html,
The more important thing is that I've learnt
the game and I can now drive it. Thus, if
the results below happen to be invalid or not
like we want them to be they can be reimplemented.
The solution is very much in the spirit of HyperLatex
that is I do not reimplement TeX. I only need
an extra mode-stack. At each recursion level
the stack informs about the mode the process is in.
The modes, at the moment, can be: vertical,
horizontal or paragraph. The definition of
an environment must specify whether and how
the stack is to be manipulated. For example,
itemize enters a higher recursion level and ifequal
(defined and used internally by HyperLatex) does not.
I follow TeX's view of paragraphs. A \par command or
an empty line _ends_ a paragraph. A visible character
opens one. The latter may depend on the surrounding
context. For example, it seems feasible not to
open a paragraph at the beginning of an item.
Lot's of things remain to be done. XML insertions
pose a little bit of a problem as they are not processed
by the standard algorithm. There should be a way of
telling whether and how an XML insertion manipulates
paragraphs. I want to focus now on this isssue and
Unfortunately, I only may spend an hour or two on playing
HyperLatex and I estimate that the full job requires some
I submit the current sources to Tom, in a private letter.
\item C\par\par D\par E%
\item G\par\par H