6

As suggested in, e.g., this question, I use non-breaking spaces in LaTeX to avoid some bad line breaks, e.g., "the dimension of~$X$"

However, in some cases, this leads to words being hyphenated in an unpleasant way. For instance, when writing "Let $E$ be a vector space of dimension~$d$", everything is usually fine, but sometimes LaTeX will decide to hyphenate "dimension". Yet I find the line break "of dimension / $d$" less jarring than the line break, e.g., "of dimen-/sion $d$".

I fix examples by hand when they occur, by removing the non-breaking space; but this is unpleasant because it needs to be added back whenever the text changes.

It seems to me like the correct solution would be to have an "almost-non-breaking space", i.e., I would like to be able to write "Let $E$ be a vector space of dimension#$d$" with some character or macro #, with the effect that the space should not be broken except if it would cause a hyphenation. Is there a reasonable way to achieve this?

[As TeX has a way to tweak the penalty associated with hyphenations (see, e.g., this answer), I would imagine that there ought to be a way to specify that a line break at one point should be possible but discouraged unless a word needs to be hyphenated... hence this question.]

  • There is always Let $E$ be a vector space of \mbox{dimension}~$d$, but that, of course, makes appropriate hyphenation harder. In such cases, sloppypar is an option. – Steven B. Segletes Dec 22 '16 at 18:41
  • @StevenB.Segletes: Thanks, I hadn't even thought of this! It's a good idea, but as you pointed out this will make LaTeX mess up in cases where hyphenating "dimension" is the only reasonable solution... so it's not ideal. – a3nm Dec 22 '16 at 18:47
  • 2
    Try dimension\penalty1000\ $d$. This means that the space is a viable breakpoint but breaking will cost a penalty of 1000. A penalty of 10000 would correspond to ~. (My advice, don’t try to outsmart the system) – Henri Menke Dec 22 '16 at 19:10
  • 1
    @Werner: Yes, thanks for your point, but the hope would be that such a macro would reduce the number of cases where manual intervention would be required. I don't like leaving stuff for "end-of-document modifications" because one always finds more "last things" to change to the document in the end... – a3nm Dec 22 '16 at 19:10
  • 2
    @HenriMenke: Thanks! After testing, this appears to work fine with a penalty of 100; starting at 101 LaTeX seems to prefer to hyphenate a word. Then I guess the easiest way to use it would be, e.g., to \newcommand{\nb}{\penalty100\ }, right? If yes, please feel free to post your comment as an answer! – a3nm Dec 22 '16 at 19:26
6
   a\nolinebreak [3] b

Has a normal space and discouraged line break

  • Thanks! From trying it out, this works, but the value should be [1], not [3], otherwise LaTeX still prefers to hyphenate the word. As per tex.stackexchange.com/a/94220 this is consistent with my experiments with @HenriMenke's solution in the comments to my question. – a3nm Dec 22 '16 at 22:41
  • @a3nm [1] is a penalty of 51 so hardly any discouragement at all (51 in a range 0-10000) , it will be almost always do the same as just using a normal space, but if it does what you want then that's fine:-) – David Carlisle Dec 22 '16 at 22:45
  • 1
    however downvoted, care to leave a comment saying why? – David Carlisle Dec 22 '16 at 22:47
  • 2
    @a3nm: Elaborating on David’s comment, it’s true that \linebreak[1]\space charges a “penalty” of 51 (arbitrary units) for breaking the line at the \space, but, on the other hand, a break at a hyphen (inserted during hyphenation) is charged 50 units under the conventions usually adopted by LaTeX. Thus we see that [1] does correspond to the minimum penalty value that will make TeX prefer the break at the \space over the break at the hyphen, if this does not otherwise affect the “badness” of the lines being formed (actually, the total demerits, but this would be a pretty long story…). – GuM Dec 22 '16 at 23:22
  • 1
    @GustavoMezzetti yes ~ is like \nobreak\ rather than \nobreak\space so ignores spacefactor – David Carlisle Dec 22 '16 at 23:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.