Amdahl UTS vs. Unix/V and Berkeley 4.2

[lacasse at RAND-UNIX.arpa]

We evaluated UTS for quite a while here, back in 1982 or so.  This is my
personal opinion, and not that of Rand.

The biggest problem we had is that our IBM hardware didn't have full
duplex tty lines.  We wanted to use screen-oriented software (both that
did, and did not use the curses library).  We tried leasing a fancier
full-duplex terminal controller, but it was very expensive.  One fellow
here hooked up two tty lines to one terminal, and ran the Rand Editor on
it that way.  This was kludgy, and very expensive per-terminal as well.

The file system had a fragment size of either 4 or 8Kbytes (I forget
which).  I thought that was a little wasteful of disk space.

It had some unusual conventions, like a standard directory in everyone's
home directory called "...", where .login, .cshrc, .profile, etc.
ad infinitem were located.  This is a fine idea, but I'd rather AT&T
or Berkeley made such major inovations.

I think it had problems with super-recursive programs (stack overflow).

The executables, especially a minimum executable (compiled a.out of
hello_world.c) were ususually large.

It was quite fast, and had good floating point and integer benchmarks.

The most telling conclusion is that we stopped using it, and bought
a Vax 11/780 for that application (the one we had tried using UTS for).

They may have made dramatic improvements since then.  I'd advice you
pay careful attention to the full duplex tty issue.

      Mark LaCasse                  qantel!hplabs!sdcrdcf!randvax!lacasse
      c/o The Rand Corporation       cbosgd!ihnp4!sdcrdcf!randvax!lacasse
      1700 Main Street              lacasse at Rand-Unix
      Santa Monica, CA 90406
	213/393-0411  ext. 7420