[REBOL] Re: another piece for pdf-dialect
From: dness:home at: 30-Jul-2001 15:26
Gabriele Santilli wrote:
> Hello Robert!
> On 30-Lug-01, you wrote:
> RM> Hi, AFAIK does TeX use such a multi-box approach. Each
> RM> character uses a defined 2D space, depending on the font
> RM> settings. TeX sums all those boxes and knows the size of the
> RM> page, therefore it should be possible to calculate a page
> AFAIK TeX tries not to break paragraphs, avoids putting section
> titles or spaces at the end of the page, etc. Ita has a lot of
> nice euristics to create the best layout.
> RM> break. Do you have access to Donald Knuth's book about TeX?
> I don't, unfortunately. I've not been able to find Knuth's books
> in local book stores or libraries...
> RM> The basic algorithm isn't that hard I have heard. And let's
> Indeed, as I said, a QAD solution wouldn't be difficult to add.
PMFJI but I have a developing interest in REBOL and considerable past expertise
with TeX. I have missed the discussion up to this point, but I will go back
and see if I can re-construct it and perhaps add something about `how'
TeX does `whatever' you are interested in. In the meantime, I would be
happy to hear (and try to answer) specific questions.
To discuss a point raised in this particular communication, (I hope I have
the context right), TeX does `paragraph breaking' through a rather elaborate
set of `penalties' which are evaluated using a dynamic programming algorithm.
This allows the TeX writer to describe his/her `feelings' about the things in
typesetting that are `not nice'. For example, a `title' line as the last line
of a page is _very_ `not nice'. Similarly, paragraph spacing may be quite `strechable'
but if you make it either way too big or way too small, you have high penalties.
I hope this is relevant to your conversation...