What to call \&, redux (was: warning on mid-input line sentence endings)

[G. Branden Robinson]

Hi Peter!

At 2022-07-17T11:47:59-0400, Peter Schaffter wrote:
> > Note the omission of the ineffable \& escape sequence after one of
> > the initials.
> 
> Appreciation for the exquisite choice of adjective. :)

I've adopted it in emails since I started to get yelled at every time I
used my own nomenclature for it.  Also, this is shorter than all
proposed alternatives.  :D

> O.T., I know, but have we reached a consensus on how to make
> that effing \& effable?  I'm still for "non-printing, zero-width
> character."

Modulo a hyphen, that's what Ossanna had in CSTR #54 (1976), and it's
damn good.  It says nothing misleading.

The main thing giving me pause is that, in groff, it also perfectly
describes `\)`.  They're both dummy characters, but \) is transparent
to sentence endings (all of a sudden we're back on topic) and \& is not.

As an amateur documentarian and lexicographer, I have a powerful
hankering for definitions I can capture as noun phrases without having
to resort to relative clauses.

But these escape sequences have so far defied my efforts to characterize
them tersely.  The docs in Git HEAD still call \& a "non-printing input
break" but I'm not wedded to "input break", or even to "non-printing"
for that matter, despite its Ossanna stamp of approval.  I'm strongly
opposed to calling either of them a "space".

In my random walk along the mountainside of definitional perfection, I
hit one waypoint I liked.

\&      zero-width sentential dummy character
\)      zero-width non-sentential dummy character

But I still have two reservations about it.

1.  The distinction isn't really whether the character is constitutive
    of a sentence, but as noted above, whether it will maintain a state
    of sentence cessation.

2.  I am dubious that I can sell the word "sentential" to most readers,
    even if it were perfectly accurate in this case.  I once had a
    rather chagrining experience at a former employer, casually dropping
    this term into a defect report and finding that I had startled the
    entire internal application support team with the fact that this
    word even existed.  While I'm in confessional mode, I'll admit that
    I provoked a similar response when filing a ticket pointing out that
    some operation in the system didn't conform to Peano arithmetic.

    One can seize the nerd crown only to find that it doesn't wear
    comfortably...

The qualifier "non-printing" might still prove valuable for cluing in
the reader that an escape sequence is effectively eliminated early in
processing; in other words that it doesn't "become" a "zero-width
space", which many readers seem to experience a strong temptation to
believe.

So, yeah, still stuck on this issue, and that's why.

Regards,
Branden

signature.asc
Description: PGP signature