Re: [NTG-context] Fwd: [XeTeX] Redefining the footnotes in XeTeX

Hi Hans, Am 21.11.2007 um 14:02 schrieb Hans Hagen:

Steffen Wolfrum wrote:

...

BTW: are the benefit's of David's work also available for ConTeXters?

i've never looked into it, but because latex and context are set up completely different it probably would mean more refactoring than i'm willing to do

concerning multiple footnotes, it's been there for a while

What Brian, starting that thread on XeTeX-list, ment was this: Am 20.11.2007 um 15:32 schrieb Brian Anderson:

What I'm trying to do is change the footnotes at the bottom of the page from this:

1) first note 2) second note .. 6) sixth note

to this:

1) first note 2) second note 3) this is a long footnote, so it is on its own line 4) short note 5) short note 6) short note

I fear this is not (yet) possible in ConTeXt, right? Anyway, in the meanwhile there was a verbose answer to this:

Was" Datum: 21. November 2007 14:27:11 MEZ Betreff: [XeTeX] Redefining the footnotes in XeTeX

Hi

If you are using plain XeTeX rather than the LaTeX variety, you might be able to tweak the example on p. 398 of _The TeXbook_, perhaps by measuring each footnote at the start of your definition of \fn and then if it is over a certain percentage of \hsize putting it all in an \hbox of length \hsize:

\hbox to \hsize{Text of longish note.\hfill}

That should, I think, force the relevant notes to occupy a line by themselves. But it might cause formatting difficulties with a preceding line of short footnotes, unless you could add \hss at the end of each footnote. Anyway, perhaps worth fiddling with. (If you have any notes that occupy more than one line, you would need to set the calculation _not_ to do the above trick since it would overfill the \hbox and in any case you would presumably want any following short note to appear in the breakline of the long note.)

My own ploy for doubling up footnotes (currently only allowing automatic setting of a short footnote full right in the breakline of a previous footnote, though it could in principle be made into a more sophisticated routine, I think) is to do various measurements every time a footnote is encountered to see if there is enough room (the space available gets put into a dimen \availablespace). The whole of the \fn command is given as an

\if \else \fi

sequence (since if the footnote is over a certain length it can't be doubled up with the previous note anyway), and I use a double set of braces round footnotes since I might want to tweak the footnote arrangement manually (either because it has messed up or because it happens to look better on a given page - e.g. to reduce white space between text and notes - if a short note does after all go into its own line). So:

{\fn{This is the footnote.}}

If I want to make sure that that footnote is _not_ doubled up (even if there is room to do so), I can give

{\availablespace 0pt \fn{This is the footnote.}}

And if I want to ensure that it _is_ doubled up even if (for some reason) it isn't doing it, I can give

{\availablespace \hsize \fn{This is the footnote.}}

Now, to calculate the \availablespace on completion of each note (so that the next note knows what it is), my version of \fn puts the note into a temporary \hbox and then sets up a \vbox which contains the \unhboxed text of the note surrounded by $$ $$. That lets me get at \predisplaysize and I can use that to make the calculation, which it does by setting \availablespace globally so that the measurement migrates out of the group in which all this is going on, and can be used by the next note.

I would post the entire macro but it's embarrassingly ropey, written many years ago when I first made acquaintance with TeX and was asked by the publisher to double up the footnotes. It has unresolved glitches (hence the double braces {\fn{}} which allow local control immediately before the \fn) - in particular, it doesn't interact well with the page-breaking mechanism so that it will place a short footnote full right even on a new page (I override this by giving \availablespace 0pt). I've never got round to attempting an output routine that would ensure this doesn't happen, though it should certainly be possible.

Anyway, perhaps a few ideas for you to ponder over your evening cocoa...

Best

John

Hans, can you build a routine upon this for ConTeXt? (Just in case it's not too weird!!!) Steffen