Re: Groff info manual section Section 5.1.3 Sentences should be changed to also mention that end of sentence characters are recognized when followed by two spaces

[G. Branden Robinson]

At 2020-12-02T16:10:52-0500, T. Kurt Bond wrote:
> I think that the treatment of sentences in this new version of the
> groff manual is much improved.
> 
> An oddity: In the PDF version, the lines that purport to show the
> output results of the Bertrand Russell example only do if you are
> using -Tascii.  -Tutf8 and -Tpdf actually show the curly quotes that
> \[lp] and \[rq] produce for those output devices.   I have no idea if
> this minor discrepancy should be mentioned in the text.

This was deliberate, and perhaps one could argue "sneaky", on my part.
When Texinfo renders to regular Info it produces ISO 8859-1 (Latin-1)
output.  So no typographer's quotes will be available.  When rendering
to TeX (and ultimately DVI/PS/PDF), typographer's quotes _are_
available.  Similarly, whether \[lq] and \[rq] produce typographer's
quotes or simply " depends on the groff output device, and as a rule the
examples in our Texinfo manual are not device-specific.

After groff 1.22.4 was released I started using a Texinfo 5.0 feature
that renders "code quotes" (such as those in examples) as they appear in
ISO 8859 and Unicode charts.  That is, ' is a neutral apostrophe and `
is a grave accent.  Doing this improves the appearance of the (PDF form
of) our Texinfo manual considerably, and also makes it clear what is to
be typed.

So the choice of escapes and sample output in my example is a subtle way
of introducing the reader to the important notion, regarded with fear
and loathing by some man page authors, that What You See May Not Be What
You Get in groff glyph rendering.  Say what you mean and trust the
output driver to make its best effort to fulfill your request, and throw
a warning diagnostic if it cannot.

Getting into special character escapes and device-specific glyph
repertoires is premature at that point in the manual, and it would be a
major digression to explain that detail of the example in the text, even
in a footnote.  My best attempt at a complete and frank discussion of
these issues is in the groff_char(7) man page[1].  Eventually I'd like a
lot of that material (but not the glyph tables) to be reflected in the
"Using Symbols" node of our Texinfo manual--but that's a post-1.23.0
ambition.

Sometimes in technical writing it is useful to permit the reader to
observe the presence of deeper waters without calling attention to them.
Whether I chose a good opportunity to do so in the instant case is, of
course, subjective.

What do you think?

Regards,
Branden

[1] A version not too far out of date is in rendered form at Michael
    Kerrisk's man-pages project:
    https://man7.org/linux/man-pages/man7/groff_char.7.html

signature.asc
Description: PGP signature