> From: Gregory Pittman <gpittman at iglou.com> > > On 04/11/2014 08:17 AM, Peter Nermander wrote: >>> I'm sure you are right about the "original" meaning of non-breaking >>> spaces. Let >>> me say it other words: It would be nice to have two different kinds of >>> spaces: >>> one with fixed width and a streching one. It doesn't matter if you call it >>> >> >> There are plenty to choose from: >> >> https://www.cs.tut.fi/~jkorpela/chars/spaces.html >> > I think the important question is about what happens to a non-breaking > space with full justification. Is it subject to being widened? If so, I > think Scribus should exempt these from any stretching. >
I think we need a nonbreaking space that stretches when justified. If you really want a fixed width space, use one of enspace, emspace (1,1/3,1/4,1/6) or figurespace. In the future Scribus will emulate those (and other spaces) if they are not present in the font. Currently my idea is to make only spaces which are defined as a multiple of a quad or have "non-breaking" in their name non breaking and all others breaking. For a sequence of spaces the last one would determine if there is a break or not. Another idea is to implement french spacing in the layouter. In that case the layouter would convert any space, nbspace, thinspace, narrow no-break space or missing space to the defined French thin space in positions where required. To avoid this space in special cases you'd have to use zero width no-break space or one of the fixed spaces. In any case there would be no break position. /Andreas
