Lars wrote,

 > At 21.35 +0100 2000-03-05, Reinhard Kotucha wrote:
 > >Hi,
 > >unfortunately Frank told us in his last mail that he vanishes for a
 > >few days to work on output routines.
 >
 > I wonder when he will appear again? It's been a month. OTOH, output
 > routines are tricky things, so it is hardly surprising that it takes time.

here he is but he is going to vanish again in a few minutes :-)

have to do preparations for the DANTE meeting where i like to show some of the
new stuff

yes output routines are tricky beasts but the work we did last mont was well
spend, i think. i hope to be able to present parts of it real soon on this
list on the www server.

so much for now:

 - the pages produced from the new OR are looking horrible :-)
   and it is slower and can't do a lot the old could
 - but we are really feeling good about it as it has the potential
   and it is getting there.

what it can right now is

 - doing multiple columns (up to 6 right now but this is not a restriction)
 - handle different float strategies, eg float after callout, float after or
   on the same column, float on the same page
 - deal with bottom floats  when footnotes are involved (disallow them, decide
   where the footnote goes ...
 - allow absolute manual control if needed. ie writes a float position file
   and all you have to do is to move lines in that file to move floats around.
 - allows for \flushfloat which outputs all floats before the current point
   without starting a new page
 - allows for caption positioning depending on placement
 - allows for changing the position algorithm

 - will soon handle page spreads (essentially considering a spread as a
   multicolumn situation)
 - will soon do spanning floats
 - will (hopefully) soon do marginal floats

right now it is roughly 40 pages code and growing (and buggy :-)
but ...

as for the page hook. somethig like that should be definitely become part of
that OR.

 > PS: When I wrote about tracefnt anyway, I might as well mension this thing
 > too. Could it be possible to have the option reinserted that temporarily
 > resetted the \tracing... parameters during \selectfont? Appearently this
 > feature was bundled with something else that probably doesn't work anymore,
 > but I can't see any reason (apart from lack of time) why it shouldn't be
 > possible to reinsert just this particular feature. I'm so tired of wading
 > through screenful after screenful of font selection every time I need to
 > trace a piece of code that, apart from the things I'm really interested in,
 > also does a couple of font changes.

wish granted. in fact while working on the OR and other things i got so tired
of such output that i wrote a small little package called trace that
implements that kind of stuff. putting it into tracfnt is not a good idea as
there is more than just font stuff one wants to disable.

it is right now in Barbara Beeton's hand as a submission for TUB and i'm going
to put it onto the WWW in a few days.

good night
frank