Re: [Groff] rationale for italic correction mechanism?

[Dave Kemper]

> There is no conflict to groff's `inability' to apply it automatically.
> What you really want is a *macro package* which does the job.

Well, having a macro package which does the job is certainly better than
not having it.  And thanks to those who have pointed out macros which
do the job: I can poach some of their code even if I'm not using that
particular package.

> This is
> exactly the same as with TeX and LaTeX: While TeX (the program)
> doesn't apply italic correction automatically, there is LaTeX's \emph
> command which does it.

I have complete ignorance of TeX/LaTeX implementation.  But using
parallels in the groff world, it seems that groff handles most
typographical issues at the character level (kerning, ligatures, etc.),
whereas macro packages tend to take on issues at the layout level
(margins, paragraphs, headers, page numbering, etc.).

To me, italic correction's closest kin is kerning: both are adjustments
to the horizontal space between glyphs that are made to correspond to
widespread practice in professional typesetting.  Yet groff handles them
in fundamentally opposite ways: kerning is a global setting, done by the
basic typesetting engine, and turned on by default; italic correction
must be done case by case, can only be made global with use of a macro,
and is not done by default.

In addition, while the problem *can* be solved with macros, it's a curious
problem to *require* macros to solve, as it's the rare case indeed
where one would want to leave the characters in question overlapping.
In general, groff uses sensible defaults for typographic issues.
Italic correction is a curious exception.

So perhaps a more accurate phrasing of my question is, why are two
functions as similar as kerning and italic correction handled in such
different ways?