[TUHS] Curly braces: An evolution of UNIX and C

[Phil Budne via TUHS]

Regarding roff, 4.2BSD roff and v7 had "device" support by reading an
a.out file (with an initialized struct) with spacing and character set
info, pretty much oriented to PRINTING devices (Diablos and such)
TTY37 was definitely one of the profiles, "lp" was another.

The first Unix system I saw was an 11/40 in the EE department
of Stevens Tech somewhere in the 79-81 time window.  The console
was a DECwriter I (LA33) and there was one (or more) VT05's
(UC only, I swear they must have been made from cast iron).

In the same time window the computer center ran a DEC-10.  Only the
two full-time programmers had VT100's. The terminal room had one VT52,
four Datamedia Elite 1500's (with APL support), and a mess of ADM-3A's
and DECwriters (III's?).  I think a populat terminal on ITS at the
time was the Datamedia Elite 2500.

I started at DEC in 1981, and we all got VT100's in our cubes, BUT
take home loaner terminals were VT05's at the time (ooof!), and not
TOO long after that we got VT52's.  A friend from a Cost Center with
more money to burn got a home VT100 not too long after that...

So while the ECMA 48 may have come out in 1976, the VT100 in 1978, and
the ANSI standard in 1979, it took a NUMBER of year before the it was
the dominant paradigm, and many of us were dealing with the fact that
just about ANY VT100 escape sequence put a VT52 into "paged" mode, or
some such dain bramage.

Regarding any idea that "vaguely ECMA-48-compliant video terminals"
were always the majority, or "the DEC VT100 showing up and commencing
to eat the world", there was a (seemingly) long interval where
terminals were a pretty unrully bunch thru the 1970's and early 80's,
and it's little wonder that termcap was invented to tame them.

To bring things firmly back into TUHS territory: In the early 80's I
took a Harvard Extension class, and remember the Harvard Science
Center 11/70 UNIX systems; ISTR them as being v6 based, and terminal
types were all two characters (presumably passed around as two bytes,
without the benefit of environment variables).  And that termcap had
two-char aliases as well... Am I remembering this correctly?