Subject: Re: People only buy support for bad software ?!?
From: "Stephen J. Turnbull" <>
Date: Thu, 3 Jun 1999 15:34:59 +0900 (JST)

>>>>> "Brian" == Brian Bartholomew <> writes:

    Brian> I prefer to say that while some of the bread crumbs are
    Brian> missing, I can still follow the trail.  There are good
    Brian> business, legal, and reputational reasons why I don't
    Brian> expect anyone here to admit publicly to deliberately
    Brian> writing software poorly.

True.  Especially since it's almost surely not true that anyone here
is deliberately writing poor software, so they're also under an
ethical obligation not to increase FUD by saying that they do.

So, why hasn't anybody mentioned "Windows NT Workstation" yet?  Just
because Microsoft wrote "Server" first and then crippled it to create
"Workstation" doesn't mean that they didn't intentionally create
software with less functionality than they knew how to.  Microsoft
admits _that_.  Any man-on-the-street can see that even if they didn't
do the registry thing, it costs them nothing---_production_ cost---to
replace the "Windows NT Server" label on some of the disks with a
"Windows NT Workstation" label.

Exactly the same argument applies to Word CDs vs. Office CDs.

In fact, anybody who ships media with less than 640MB of bugless
program per disk is guilty by Brian's criterion as far as I can see.
(Straw man, of course; but I don't see Brian making any attempt to
draw lines that would exclude that interpretation, either.  He has
explicitly included omitting functionality as "bad practice.")  But
this is just plain bloody-minded.

_This kind of versioning is a socially good thing_ (with the proviso
that Microsoft's profits be appropriately taxed and the revenues
returned to the appropriate customers).  In the usual economic sense
that all customers and Microsoft are simultaneously made better off
than charging a uniform price for all units.  (The FUD and outright
lies told by MS officials are not, of course.  I'm just talking about
the "reprehensible" practice of charging different prices in different
transactions, here by providing "different" products with the same
production technology at the same cost.)

Another eg: releasing a partially functional product early is clearly
a good thing if so labelled (as "partially functional," not "good
thing").  The developer knows how to make the better software, just
hasn't done it yet.  (Same Ramsey pricing, although it takes some
twisting of the usual definitions to get it to come out, and you won't
find that interpretation in the textbooks.)  If freely available for
download, and all improvements will be, I see no ethical problem here.

I'm not going to discuss the analysis on FSB.  It's a theorem, proved
in 1926 (IIRC), but the bandwidth-to-education ratio on this topic has
been near infinity in the past---so I refuse to argue.  See Shapiro
and Varian's _Information Rules_, chapter on "Versioning" for the
mechanism in the context of software.  Then any moderately advanced
industry economics textbook for "Ramsey pricing" in general.  Also
"price discriminating monopolies."  A good one will explain the
relationship; a side-by-side comparison of the pricing formulae should
be enough for you though.  A very poor web version is available at

It will be upgraded over time; for now I really recommend the print
books (for one thing, I don't think the formula is in there and I'm
quite sure it doesn't even mention Ramsey yet---I just thought of that
now.  Mea maxima culpa ;-)  BTW, the taxation proviso does not make the
result trivial, nor is it necessary (for economists; non-economists
have _justified_ trouble accepting the idea without it---economists
study the appropriate caveats so they can go without saying.)

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
What are those two straight lines for?  "Free software rules."