After some private discussion with Morten on some ideas that began on
c.t.t., he thought it would be a good idea to actually not talk
amongst ourselves (how conspiratorial) but rather, talk about them here.
First on the list: is it possible to get CVS read access to the
LaTeX3 repository? I was thinking vaguely about looking things over
and adding to the documentation (since, last I checked, it was rather
sparse) as a good way to accustom myself to it all, but I'd prefer to
not have to manually sync with your changes manually over the web
(I'm a little wet behind the ears with regards to CVS, I'm afraid.)
Secondly, I asked how xor was going, since last I heard there was
algorithm trouble with the way multicols balance with the existance
of remaining floats. The reason I bring it up is 'cause the problem's
been going around in my head recently, and I've got a few very half-
baked ideas -- but it wouldn't be productive for me to do anything
about them if that particular problem had already been solved.
Finally, somehow the topic of extensions to pdf-e-TeX came up, and
there were some naive comments from me and some interesting comments
from Morten. I proposed the idea of something like \previouschar,
which we interpreted in two separate ways:
- it "would return the slot number of the most recently seen char
or ligature node independently of whether we just saw a glue node or
another char/lig node; any other node type could reset it to
- it would actually be more like LaTeX's \@ifnextchar, and look at
the input token list [me]
Implicit in my thinking was that you could also eat up previous chars
in the same was as \unskip with a sort of fine-grained \lastbox; to
which idea Morten replied:
This is exactly what some languages need.
Since there are so many different node types the only safe
(IMO) way is to backtrack one node at a time so ideally we
would want the pair \lastchar and \unchar. And then we might
as well get one for each different node type. I guess someone
has to come up with a good idea about what to do with \discretionary
because I'm not at all sure what the semantics should be.
Obviously, he's thought it through a lot more thoroughly than I have.
(Ah, too many "th*ough"s!)
While it is true that this idea is required for some sort of TeX-
level approach to typesetting for various languages, I was thinking
about it from a more generic level; XeTeX has shown us that it might
be a lot easier to leverage the work of other people for unicode and
crazy font contortions than to re-invent the wheel again...
The *downside* of not re-inventing is that we're stuck with whatever
we're given. For example, in XeTeX's case, obscure multilingual
typesetting is impossible for now because the shaping algorithm for
OpenType fonts for most complex languages simply hasn't been done in
the library he's using. And the alternative, AAT fonts that shape
themselves, is an obscure Apple invention that never took off in the
market for several reasons...
I guess that's all I have to say for the time being. Any answers,
reactions or opinions?