Re: [Groff] [off] micro-typography

[Sigfrid Lundberg, NetLab]

On Fri, 8 Feb 2002, Werner LEMBERG wrote:

> Increasing/decreasing interletter spacing and using horizontally
> extended/condensed font instances, both in a very conservative way so
> that it isn't visible.  The goal is to improve interword spacing to
> be as uniform as possible across the page.

I read Han The Thanh's thesis. It was very interesting, and well written.
The story about Gutenberg's bible, and how various software packages how
various software packages have been built to emulate the quality of that
piece of art, opened my eyes to thinks I've never thought of before.

In all the examples included, I have to say that this particular
aspect (extending and condensing fonts) was the one that improved quality
the least. And that was also the Thanh's own conclusion from interviews
with people looking at printed samples.

What he calls "Level 2 character protrusion", where somehow the character
protrusion was used to calculate the breakpoints, was the one thing that
has most positive effect.

In the discussion, Than quotes comments by Knuth on the implementation.
One thing Knuth sais was that he didn't like the "character
protruding" concept but would have implemented this through define margin
characters and introduced a margin kerning into the TeX kerning tables. Or
that was how I understood what he said.

Sigge