[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: DOCBOOK-APPS: jade/jadetex: section headings and rubber lengths
I've reported before on jadetex not respecting "keep-with-next" for <sect2> section headings. I _think_ I see what's going on, but I'd be grateful if anyone can confirm or deny the diagnosis. From inspection of jadetex.ltx, it looks as if the only effect of specifying keep-with-next for a section heading is to assign a negative TeX penalty to a page break prior to the heading, thus: \def\QueryPageBreak{% \ifBreakMe \ifvmode \penalty \@M \else \@bsphack \vadjust{\penalty \@M}% \@esphack \fi \else \ifnum\KeepWithNext=1 \penalty-600 \else \penalty \z@ \fi \fi The "-600" encourages TeX to break the page prior to the heading, which will have the side-effect of averting a break between the heading and its associated text. But this is obviously a bodge: what we need is not just encouragement for a break prior to a heading, but an absolute prohibition on a break immediately following a heading. I tried the experiment of copying and pasting the above fragment into my jadetex.cfg then playing with the size of the negative penalty. By making it much bigger (e.g. -6000) I could prevent some of the unacceptable breaks, but at a high cost: now a lot of pages showed breaks prior to section headings (and hence short pages) where the breaks were not in fact necessary to prevent widow headings. It seems like the -600 penalty is set a level that will (a) not do too much damage in general while (b) producing a keep-with-next effect under favorable circumstances! I'm well aware that Sebastian Rahtz is a much cleverer TeX hacker than I am, so I presume that if he didn't write a proper keep-with-next routine for jadetex it must be extraordinarily difficult to do so -- which is bad news indeed. One related observation. The text that is breaking unacceptably is admittedly a difficult one: I have a lot of short <sect2>s in succession, so keeping section headings "with next" is likely to produce short pages -- but the problem would not be so intractable if the spaces before and after headings were rubber lengths, as in LaTeX. So far as I can tell, in jade/jadetex these lengths are rigid. At a guess, there seems to be some sort of provision for rubber lengths in jadetex, via the mechanism of (sbNom, sbMin, sbMax) and (saNom, saMin, saMax), but I can't find a way of activating it; the "Nom", "Min" and "Max" values are all the same for any given section in the jade-produced tex file. Is there in fact a way (at dsssl level I suppose) of making such lengths rubber? -- Allin Cottrell Department of Economics Wake Forest University, NC
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC