Difference between different unix versions

[Doug Gwyn]

In article <17641 at adm.BRL.MIL> hxn8477%njitx.decnet at njitc.njit.edu (NJITX::HXN8477) writes:
>I know this is a stupid question, but I'll ask it any way because I am
>VERY new to unix.  As far as I know, there are 3 flavors of unix: BSD 
>which is public domain, System V which is moderately priced and Xenix
>which is astronomically expensive.  Just what is the difference between
>the 3 flavors.  And if all of them can run any unix program, what makes
>one buy Xenix, when they can get BSD?  

UNIX is a trademark of AT&T.  All three variants you mention, and several
others available from a variety of vendors, are derived from AT&T-owned
source code and are distributed under license from AT&T.  Xenix is the
trademark for MicroSoft's version, BSD stands for "Berkeley Software
Distribution" and is available in source form to UNIX System source
licensees only (version 32V or later) for a moderate fee from the
University of California at Berkeley, and UNIX System V is AT&T's name
for their commercial versions of the UNIX operating system.  Currently
source licensing seems to be available only for UNIX System V Release
3 and costs something like $65,000 (may be higher) for commercial use.

Binary releases of UNIX are prepared by vendors working under explicit
sublicensing agreements with AT&T.  AT&T's binary sublicensing fees are
sufficiently low that end-user prices for a binary UNIX can be on the
order of a couple of hundred dollars or less.  Binary distributions,
until recently, need not conform to any particular standards and so many
different variations are provided under the name "UNIX".

UNIX is emphatically NOT in the public domain.  It is AT&T property and
is protected primarily by "trade secret" provisions.  Enhancements made
by third parties are generally their property.

There are a few UNIX "work-alike" systems commercially available that
are not derived from AT&T source code.  Availability of these is of
course controlled by their owners on their own terms.

Generally, different versions of UNIX offer different features, or the
same features provided in different (often incompatible) ways, so true
binary program portability is not attained.  Source code can be designed
to successfully compile and link on a wide variety of UNIX variants, and
recent efforts at standardization (notably IEEE Std 1003.1-1988 and the
forthcoming ANSI C Std X3.159-1989) should help improve the situation.
AT&T also as of UNIX System V Release 3.2 had merged Xenix and UNIX
System V into a single version, to henceforth be marketed under the
UNIX System V banner, and have a current effort to merge SunOS (which
is based on 4.2BSD) and UNIX System V into UNIX System V Release 4.0,
at which point the features of the major UNIX variants will all have
been combined in a single AT&T UNIX System V product.  AT&T has also
been pushing for binary standards (on a per CPU family basis) to allow
future successful marketing of "shrink-wrapped" software for UNIX-based
systems (without the customer having to worry about variants of UNIX).

However, IBM has been pushing their own variant, AIX, which was adopted
by the so-called Open Software Foundation, and DEC's Ultrix seems to be
continuing its own separate evolutionary path.  Probably other vendors
will maintain their own weird flavors of UNIX into the foreseeable
future.  This clearly cripples the whole UNIX binary software market,
but what does one expect, given the short-sightedness of the MBAs
typically in charge of marketing?

The academic/research world also seems to be pursuing their own UNIX
evolutionary directions via 4.4BSD, MACH, 9th Edition UNIX, etc.  I
have no complaint against this, after all it IS research; but some
commercial vendors are likely to try to track the research directions
in their commercial products, which would also increase fragmentation
in the marketplace.

If this sounds like a mess, it is.