stevesu@copper.UUCP (Steve Summit) (04/30/86)
In "A Typesetter-independent TROFF", by Brian W. Kernighan, it is
mentioned that "Some obsolete commands have been eliminated
(e.g., .fz, .li;...)," and later, in case you hadn't noticed it,
"The .fz and .li commands are no more." I am wondering why .li
was deleted. To be sure, it was undocumented, but after I
discovered it one day while poking through the source code, I
found it quite handy.
What .li did, in case you're rusty on your obsolete, undocumented
nroff commands, was to take the next line (or next n lines, with
an argument) LIterally, i.e. without checking the first character
to see if it was a period or an apostrophe, for a request. Now,
I realize you can protect a leading period or an apostrophe with
\&, but this always seemed kind of unclean to me. The .li
request said what I meant, whereas a \& is a dodge that just
happens to work.
Is there a modern, cleaner way to suppress request processing,
other than \&? Temporarily changing the control character
with .cc is also ugly, because you have to pick some other
character, and you have to remember to set it back when you're
done.
Was there something insidiously dangerous about .li? First
undocumenting, then removing it seems a bit harsh if it was just
superfluous or redundant. There are plenty of nroff requests
which can be simulated with other nroff requests.
If you have any clues on this little mystery that you'd like to
pass along, please respond by mail, since I don't read this
newsgroup regularly.
Steve Summit
tektronix!copper!stevesu