[texhax] italic correction w/o italic [was ...anglbrackets...]

[Uwe Lück]

Am Mittwoch, den 20.03.2013, 12:01 +1000 schrieb David Crosswell:
> On 20/03/13 11:37, Reinhard Kotucha wrote:
> > On 2013-03-19 at 23:57:19 +0100, Uwe Lück wrote:
> >
> >  > while I still have not understood \/ entirely 
> >  > ("provisional" for the case of \it?)
> >
> > It's mentioned in the index of The TeXbook.  AFAIK there is an
> > italics-correction entry for each glyph in the tfm files.
> >
> > If you want to investigate, the program tftopl is your friend:
> >
> >   tftopl --help
> >
> Not being able to recall the post that mentioned application of this,
> the only time I use \/ is to assist in separation of a font that
> interferes with itself on display, generally in regard to fi, ff, or fl
> combinations.

In this case, AFAIK f{}i, f{}f, f{}l are an alternative, 
another issue that I have not understood completely 
(what is better, and why it works -- The TeXbook once 
 confused me here).

> Is that the application here?

No, ...

According to The TeXbook (p. 14), the italic correction \/ is
appropriate when changing from a slanted (e.g., italic) to an 
unslanted font, except before period or comma.

With LaTeX, you may not know about \/, \textit and \textsl insert 
it automatically. There may be cases when the automatic behaviour 
is wrong, then you can prevent it by \nocorr (if you know about 
this feature).

What I wonder about is that I sometimes find other applications 
of \/ in macros by very knowledgeable authors. I remember 
applications after boldface. The recent example was

    \def\<#1>{\leavevmode\hbox{$\langle$#1\/$\rangle$}} 
        % syntactic quantity

from Knuths manmac.tex for typesetting The TeXbook on top of 
plain.tex. It is used to typeset TeX's Backus-Naur syntax rules. 
The argument is typically typeset in \rm, I cannot remember 
having seen anything else. I wondered what rationale differing 
from what The TeXbook tells is behind such a use of \/ after 
roman (or after upright boldface).

In the special recent example, I now guess: The italic correction 
makes up for "ink" looking out of the character box to the right 
(TeXbook p. 64). The ink of a closing bracket touches its left 
box border even at its top, and it is from a different font than 
what is at the left of it. Without the italic correction, there 
is a real danger that the closing bracket touches the final 
character.

You can search the texhax archives

    http://tug.org/pipermail/texhax/

for "italic correction", you get 75 hits and a large thread 
"cases for italic correction". Unfortunately, I have not 
found a posting listed there dealing with its subject 
(it was mainly about the "d" in math italic).

Cheers,

    Uwe.