RE: flexible line-spacing?

Subject: RE: flexible line-spacing?
From: "Frank A. Christoph" <christo@xxxxxxxxxxxxxxxxxx>
Date: Thu, 13 Jan 2000 20:01:53 +0900
> If I want to have a document formatted with the pages on one spread being
> balanced (i.e. the content of the page shall fill exactly the page height
> less the top and bottom margins, which is normally done by putting the
> placement path of the last text line on the bottom margin
> itself), this can
> be achieved by having the "space-before" or "space-after" of a paragraph
> being of type "display-space" with a minimum and a maximum length
> spec. This
> works fine if there is enough "stretchability" in these specs.
> Sometimes though the stretachbility is not enough, or there is no space
> between paragraphs at all (e.g. if a long paragraph fills a page
> completely). Normal typesetting would then resort to changing the
> line-spacing a little bit. But how can this be achieved in DSSSL?
> I did not
> find any characteristic to have such a flexible line-spacing generated?
> Has this simply been forgotten in the specs, or am I missing something?

Probably not forgotten, only unaddressed. DSSSL makes a point of being
ambiguous on the treatment of line and page breaking, leaving these things
to be implementation-dependent. (I believe the introduction to the standard
explicitly says something to this effect.) For example, although the
standard mentions the minimum and maximum conditions on length-specs which
you refer to above, it says nothing about how they affect breaks. Any
reasonable implementation will of course try to make use of such
information, but you could have a conforming DSSSL implementation which
totally ignored them.

The reason for all this, I believe, is that the designers wanted DSSSL to
support a wide range of existing backends, like RTF and TeX, many of which
also have ill-defined breaking semantics. Another contributing reason is
also undoubtedly the fact that there is no universal notion of line-breaking
that applies to all natural languages, and DSSSL is pretty open-ended when
it comes to language support. So you see, ambiguity breeds ambiguity; this
is an instance of the computing world's own version of entropy.


 DSSSList info and archive:

Current Thread