3

It's pretty standard for latex to add a little whitespace between captions, figures etc to reduce the chance of page breaks mid-paragraph. In some cases I feel latex is a little too liberal with its stretching. This may contribute to the length of my document too.

I'd like some control over the threshold of the decision to page break mid-paragraph or stretch, similarly to the way \emergencystretch can be adjusted for whitespace within a line. Is this possible?

[EDIT]
Perhaps penalties is what I'm after, but I haven't got anything working so far.

Community
  • 1
jozxyqk
  • 1,089
  • the space between captions and figures is usually in a box so never stretches. Any vertical space (such as \parskip will have specified stretch and shrink limits so these can be reduced to suit whatever you want) As always your question would be a lt clearer if it had a full example document showing the problem – David Carlisle Jun 18 '14 at 12:08
  • @DavidCarlisle this sounds like its exactly what I want. Could you point me in the right direction to change these stretch/shrink limits? – jozxyqk Jun 18 '14 at 12:28
  • 1
    There are vertical spaces specified all over a typical class file, between paragraphs, before and after lists, between list items, before and after math displays, before and after section headings. It is the job of the class designer to set these in an appropriate way without knowing what class you are using and which spaces you need to address its hard to offer advice. Changing individual spaces in an ad hoc way is unlikely to make a consistent result. – David Carlisle Jun 18 '14 at 13:41

1 Answers1

1

Breaking mid-paragraphs is controlled mostly by \interlinepenalty if you set it to the maximumn

 \interlinepenalty=10000

then page breaking between lines of a paragraph is prevented.

The default value is 0 so if you set it to a negative value you increase the likelihood of a page break mid-paragraph.

TeX will only break in a different place though if there is sufficient vertical stretch to make more than one breakpoint feasible.

David Carlisle
  • 757,742