On Wed, Jul 16, 2008 at 06:34:48AM +0200, Werner LEMBERG wrote:
> > I have not used `tbl' for some time and now see the following
> > behaviour, which was not present before, I'm quite sure: [...]
>
> Source code, please.
>
> > and it's still funny: the table is rendered completely (but
> > extending into the right margin), only the inter-column space is
> > squeezed to zero, it seems. now: is _this_ intended behaviour?
attached is an example (using `ms' macros). although there
is nothing specific to it and I think in the meantime I
found out what's going on: the squeezing is caused by the
`expand;' directive when the table gets wider as the line
length, it seems: if you remove the `expand' without
altering the point size the table simply does no longer fit
on the page and is clipped.
wheras, empirically, it seems that the `expand' works in
such a way that it actually shrinks the table if it was too
wide in the first place. which sure is a feature in my
view... but apparently it shrinks the table as far as
possible without causing overlap of columns and that was
what I've seen causing the seamless merging of some entries
within the same row. question: would it not be nicer to stop
the shrinking when the inter-column spacing has dropped to,
say, one "m" space or something like that?
another remark: the `tbl' manpage states
expand -- Makes the table as wide as the current line length.
which is not quite right regarding what's happening for wide
tables as in my example: actually the table is squeezed if
it is too wide in the first place, but final table width
gets larger than the line length if inter-column spacing has
reached a zero value. Maybe one should add a corresponding
remark to the manpage, e.g.
expand -- Makes the table as wide as the current line length
if possble. For wide tables the effect might actually
be a squeezing of the inter-column spacing in the table
in an attempt to keep it completely on the page. The
table might extend into the right margin in this case.
> > should one not get some kind of warning ("table to wide" or
> > something like that)?
>
> I'm not sure that groff has such warnings built in.
nor am I :-). question is, would it be possible for `tbl' to
do a final check of table width vs. line length/page size?
and would it be difficult to implement this check/warning?
joerg
>
>
> Werner
.\"------------------------------------------
.ds ms "Wolfgang Müller-Schauenburg
.ds vdh "Jörg van den Hoff
.ds paul "Paul Maguire
.ds vin "Vin Cunningham
.ds roger "Roger Gunn
.ds ant "Antoon Willemsen
.ds nico "Nico Leenders
.ds rich "Richard Carson
.ds adri "Adriaan Lammertsma
.\"------------------------------------------
.SH
Detailed Agenda
.nr PS 8
.nr VS 13
.LP
.TS
expand;
|l l l l l|.
=
October 19 Activity Leader Room Manual Chapter/Topic
=
tbd tutorial exponentials \*[ms] tbd
11:00 course registration; packed lunch
12:00 Welcome and introduction \*[vdh], \*[nico] Lecture room
Chapter 1
12:15 Mathematical basis of modeling \*[vin] Lecture room Chapter 2
12:45 Computer exercises \*[vin] Computer room T{
Exponential functions: convolution
and curve fitting
T}
13:15 Basic pharmaco-kinetic concepts \*[nico] Lecture room Chapter
3
13:45 Blood-flow and one-tissue compartment model \*[vdh] Lecture room
Chapter 4
14:30 Computer exercises \*[vdh] Computer room Extraction, one-tissue
simulation, non-linear fitting
15:00 Coffee break
15:30 Energy metabolism and two-tissue model \*[ms] Lecture room Chapter
5
16:00 Computer exercises \*[ms] Computer room Two-tissue simulation
16:30 Discussion \*[vin] Lecture room
17:00 End day one
18:00 Evening event
=
October 20 Activity Leader Room Manual Chapter
=
tbd
=
October 21 Activity Leader Room Manual Chapter
=
tbd
13:00 End of course
_
.TE
