[TUHS] Rivers, the ragged right, and BCPL roff (was: Spooky action at a distance in line adjustment...sometimes)

[TUHS] Rivers, the ragged right, and BCPL roff (was: Spooky action at a distance in line adjustment...sometimes)

[Doug McIlroy]

> I was poking around on the Unix v7 man page archive the other day, and
> discovered to my delight that BCPL roff's request list was documented
> there. 

Yes, it's the same spec, though reimplementated; there was no BCPL
compiler for the PDP-11.

> Unfortunately, it was marked up in...BCPL roff embedded within
> a man page, which hopelessly confused groff.   

I edited the v7 manual and I believe I proofread every man page. How
could I have overlooked the misbehavior of the .li request*, which was
even then not in the reference manual's list of n/troff requests? On
closer examination, .li does appear in a condensed alphabetical list of
requests--and it appears in the source. But if Joe never described it,
it's no surprise that groff dropped it. It is totally unnecessary--a
bit of special pleading rendered obsolete by Joe's elegant invention
of \&.

doug

* .li causes the next line to be read literally, even if it begins
with a dot.

Re: [TUHS] Rivers, the ragged right, and BCPL roff (was: Spooky action at a distance in line adjustment...sometimes)

[Warren Toomey]

On Sat, Jun 30, 2018 at 10:23:57PM -0400, Doug McIlroy wrote:
>I edited the v7 manual and I believe I proofread every man page. How
>could I have overlooked the misbehavior of the .li request*, which was
>even then not in the reference manual's list of n/troff requests?
>doug
>
>* .li causes the next line to be read literally, even if it begins
>with a dot.

Slightly Unix related, .li is in the version of roff which I wrote in
a C-like language for the PDP-7 version of Unix:

  https://github.com/DoctorWkt/h-compiler/blob/master/roff/wktroff.c
  https://github.com/DoctorWkt/h-compiler/blob/master/roff/roff_manual
  https://github.com/DoctorWkt/h-compiler

Cheers, Warren