[NTG-context] Re: penalties

[Mikael Sundqvist <mickep@gmail.com>]

Hi,

On Thu, Oct 10, 2024 at 2:58 PM Henning Hraban Ramm <texml@fiee.net> wrote:
>
> Yesterday I learned:
>
> \startsetups[*default]
> \widowpenalties 3 10000 1000 0
> \clubpenalties 3 10000 1000 0
> \orphanpenalties 2 8000 1000
> \stopsetups
>
> If I want the penalty settings applied to “everything” (after defining
> the layout, I think?), then I need the setup "*defaults".

Yes, for now, since it got reset when \setuplayout was done. But we
are working on making it possible to also setup this with \setupalign,
with a few "reasonable" pre-defined setups. So, for example in my math
book I currently run

\setupalign[lesswidows,lessclubs,lessorphans,strictmath]

and that actually makes a few of these setups. It will of course be
possible to define your own set of penalties and use them instead.

>
> Since TeX uses a club against orphans (brutal!), I’m not sure what
> \orphanpenalty is for.

Wolfgang already explained this, but in short they are there to
discourage that a paragraph ends with only one word. We looked at it
yesterday at the meeting. You saw those penalties in action when you
did \showmakeup[hpenalty].

>
> I remember \twinpenalty is against doubled words (below each other?),
> and \toddlerpenalty against single letter words at the end of a line
> (why “toddler”?).

Indeed. Well, you can imagine a little toddler not really being able
to join the rest, or if you want, tied to the rest.

>
> widows and clubs get only 150 in TeX, but 2000 in ConTeXt, as far as I
> could find.
>
> Are there more penalties that make sense to set individually?

If you do math there might be more (set above by strictmath), but I
think you are fine with the ones above. Since you are finishing a
project, please tell if more problems show up. We can have another
chat where we try to solve...

>
> WRT demerits: Did I understand correctly that demerits are the “sum”
> (some formula) of badness and penalties in a line or paragraph?

A bit like that. For each possible breakpoint you get a badness value.
Then you can see it as if TeX runs over the paragraph once more, given
all possible breakpoints, with or without extra penalties, and builds
up a demerits for the different possible solutions. They can come from
badness, penalties and additional demerits (adjdemerits not to have
incompatible lines for example). Then, at the end, the best solution
(least demerits) is the one that will be used.

We are currently writing an article on the parbuilder, with and
without parpasses. You saw an early draft at the meeting yesterday.

/Mikael
___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / https://mailman.ntg.nl/mailman3/lists/ntg-context.ntg.nl
webpage  : https://www.pragma-ade.nl / https://context.aanhet.net (mirror)
archive  : https://github.com/contextgarden/context
wiki     : https://wiki.contextgarden.net
___________________________________________________________________________________