1-2 vs unlimited licenses (Unix for a 386)

[Piercarlo Grandi]

In article <16031 at vail.ICO.ISC.COM> rcd at ico.ISC.COM (Dick Dunn) writes:

   Bill Davidsen objects:
   >   I doubt that AT&T "made you do it." You say they get a higher royalty
   > for the unlimited version, why would they force you to offer a less
   > expensive version?

My dim remembrance is that the AT&T royalty is (depending on
volume) under $50 for 1-2 users and under $150 for unlimited
(complete system, I think). Their explicit reason for the not
high level of royalties was to encourage the diffusion of UNIX by
making it possible to resell its binaries for cheap, e.g. like
BellTech, Microport and Everex do; AT&T wants people to buy
UNIX/386 for *much* cheaper than OS/2, both runtime and
development system.

The reason for 1-2 and unlimited was that 1-2 was for PC users
(e.g. the ill fated 7300, or on a 386 UNIX as a competitor to
OS/2) and workstations, and unlimited was for minis and other
multiuser systems. AT&T decided to 'segment' their market this
way.

   OK, let's turn it around so you will have a harder time avoiding under-
   standing it:  They force us to charge more for a system which allows more
   than 2 users.  We have to offer licenses in two classes because:
	   - We have to offer the 1-2 user license due to economic pressure.
	     Someone who's shopping among 386 UNIX vendors, particularly for a
	     single-user system, will look at price.  If we can pay AT&T
	     less, we can maintain margin and sell the system cheaper.

Note that the entire AT&T royalty is under 10% of the 386/ix
selling price, i.e.  just a fraction of the likely margin that
ISC has on it.

If the royalty were more like 30%, as for the low cost Microport,
Everex, Belltech products, then the difference between the two
royalties would excusably be too large to ignore.

   Darryl's point is that
   ISC is simply passing along AT&T's price-tiering--to do otherwise we'd
   either have to cut into the margin on the multiuser systems or inflate our
   profit on the 1-2's.

Of course it is passing along the X Consortium's price-tiering
that forces poor ISC to sell the X development system for $795
(more than the price of the Unix development system); or maybe it
is true as somebody once wrote that ISC has huge QA costs (in
validating free sw that is running successfully off the shelf at
a few hundred or thousand sites around the world, and that is
regularly updated by the X Consortium).

	Now an interesting aside: is it possible to just compile
	the X development system off the X Consortium distributed
	sources and use it with the ISC X server? If not, why?

	Has by any chance ISC hacked the protocol recognized by the X
	server so that it is now compatible only with their
	libraries, incidentally destroying interoperability?

	Otherwise, is there any *compelling* technical reason
	to buy the X library binaries from ISC for $795 instead
	of getting the sources with which everybody is already
	quite pleased from any friend for free?

Another interesting pricing policy was DEC's with VAX Ultrix;
each extra block of eight authorized users was priced to the tune
of several *thousand* dollars, i.e. several hundred dollars per
user, up to a maximum of 64 users. Is this still true for DEC?

Is ISC going to adopt a similar pricing policy? :-> :->.

The conclusion is that if companies want to charge what the
market will bear, and/or charge high prices so as to dampen sales
because their organization is already strained under the volume
of those it makes now, they can, but please without blaming AT&T
relatively very low royalties or QA that is terribly expensive, or...
--
Piercarlo "Peter" Grandi           | ARPA: pcg%cs.aber.ac.uk at nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth        | UUCP: ...!mcvax!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg at cs.aber.ac.uk