Let's not turn this into a rerun of a flamwar...

>termcap@berkeley is dead.  termcap is generated from the terminfo master
>sources these days, and are in effect stripped-down terminfo entries
>which has the capability names and encoding of format strings translated.

Sorry, my braino.  Thanks for pointing it out.  Actually, there still
seem to be references to it in the NetBSD tree, which might have
something to do with my idiotic mistake.   otherwise, it sounds like
termcap and terminfo are largely fungible, modulo over-long capability
strings.

Last I heard, terminfo didn't have a reasonable replacement for
the TERMCAP environment variable.  I'd appreciate an update if that's
now out-of-date.

>terminfo is not limited to two-character capability names.

Personally I don't find _either_ terminfor or termcap names
particularly human-readable.  I don't see the validity in choosing an
implementation (termcap vs terminfo) based on the length of cryptic
capabilities.  As long as curses and emacs  and vi work, who cares? :)