Re: [Groff] [off] micro-typography

[Sigfrid Lundberg, NetLab]

On Fri, 8 Feb 2002 address@hidden wrote:

...

> I read the thesis too, and was also impressed.
> In fact, this inspired me to try doing the same
> sort of thing in groff, on a small sample of text
> (both character expansion and protruding punctuation).
> It worked very well, in that the evenness of layout
> was much better and there was a distinct sensation
> of smoothness in physically reading it, compared
> with the original.

One thing I've noted is that a good hyphenation file do help. Not for
creating an optically even margin, but for even out the distribution
of word spacing. I used to have a ye olde Swedish hyphenation file,
I've had it for more than 10 years. Then I picked a new when from my
TeX installation, and all of a sudden I had 1000 more patterns. Then
troff found more break point opportunities.

> Of course, this was done "with bare hands": first
> format it without these tweaks, to see where the
> line-breaks are, then go through it inserting them
> explicitly. That would be very tedious in a long
> document (such as Thanh's thesis). So there is scope
> for macro-ising it.
>
> The one person I know who has demonstrated a true
> gift for creating the kind of macro which might
> achieve it is Thomas Baruchel (he does nearly
> impossible things in diversions)!

I haven't tested his paragraph macros. Is it the ones in

http://www.ffii.org/archive/mails/groff/2001/May/0041.html

> Nevertheless, even if these effects were achieved
> to a good approximation by macros, I think there
> would still be a need to put in the finishing touches
> by hand, and any macro solution would have to allow
> this to be done.

I thought of post-processing as well. Thanh mentions work on the in
TeX as well. Someone wrote perl script for editing the dvi.

However, Than makes the distinction between Level 1 and Level 2
character protruding, where he could demonstrate that the latter was
of markedly higher quality.

       To incorporate character protruding into line breaking, the
       line-breaking algorithm should be able to take into account the
       space amount needed to protrude the characters at the margins
       of potential lines. This way, after the justification phase,
       the interword spacing will be set as the algorithm expected,
       i.e. in an optimal manner for the whole paragraph. We denote
       the use of character protruding in this case as level 2
       character protruding. [1]

While Level 1 protruding was possible in TeX using macros, this wasn't.


Sigge


[1] http://www.fi.muni.cz/~thanh/download/thesis.pdf (p. 42)