[tex-eplain] double column miseries.

Stepan Kasal kasal at ucw.cz
Thu Aug 19 09:59:12 CEST 2004


Hello,
  you requested for comments.  Here are a few, just quick ideas:

On Wed, Aug 18, 2004 at 06:20:23PM -0400, John Culleton wrote:
> \def\startcols#1{%
>     \parindent=-1em
>     \parskip=.3em
>     \hfil\bf Chapter #1: Terms and Concepts\hfil\rm\endgraf 
>   \centerline{} 
>     \doublecolumns
>     }
> ....and ended each with a macro:    
>     \def\stopcols{%
>     \parskip=0pt
>     \parindent=1em
>     \singlecolumn
>     }

I'd use grouping:

\def\startcols#1{%
  \begingroup
    \parindent=-1em
    \parskip=.3em
    \hfil\bf Chapter #1: Terms and Concepts\hfil\rm\endgraf 
    \centerline{} 
    \doublecolumns
}
....and ended each with a macro:    
\def\stopcols{%
    \singlecolumn
  \endgroup
}

> Trouble is or was, I f a single item on the right hand side required a 
> second line, then the right hand column came up longer than the left.

Well, this is easy to explain, but hard to solve.  The two line paragraph
contains clubpenalty (or widowpenalty), which says that it's not
adviceable to break page/column here.  The balancing algorithm had the
possibility of breaking after the paragraph, which would make left column
3 lines longer than the right one.

If you don't mind that the paragraph is broken, set \clubpenalty and
\widowpenalty to 0.

> And sometimes I 
> got twice the expected space after a long entry, requiring an 
> insertion of a \vskip -12pt to take up the extra space. 

That's probably because you had an index entry at the end of the
paragraph.  There appeared a line break before it.  Is my guess right?

> Perhaps my whole approach to such a listing was faulty. Indeed on 

Double column output can be difficult.

It could be possible that I'd work on an actual example with strange
behaviour if you could publish it (or prepare it).  That might improve
the eplain doublecolumn macros.

But such wrestling is always difficult, so don't expect too much.

Have a nice day,
	Stepan Kasal



More information about the tex-eplain mailing list