OS/2 Shareware BBS: 14 Text/14-Text.zip/RZIFF.ZIP/RZIFF.TXT

Text File | 1992-12-30 | 8KB | 163 lines

Robert Ziff's column in the December Corporate Computing reflects a lack
of understanding of the current OS marketplace.  Ziff appears to be
completely misinformed about OS/2, its future, and its stature in the
market.

Ziff contends:

  "The unfortunate truth for IBM is that Windows has become standard.
   The market has soundly rejected PM."

As of the end of 1992, Microsoft has shipped 18 million copies of
Windows 3.x.  Most of those copies were bundled with hardware.
Microsoft has arrangements with virtually every major PC manufacturer
(except IBM) requiring the distribution of a copy of Windows with every
machine sold.  It is delivered to every user - even those who
specifically ask not to receive it, those who already own it, and those
who have no use for it.  For this reason, Windows has a very high
percentage of "shelfware" - software that sits on the shelf unused.

Microsoft's own internal estimates (and most analysts agree with them)
are that only about 1 in 3 of all Windows copies sold are actually used,
leaving us with about 6 million copies of Windows that are in at least
occasional use.

By the end of 1992, IBM will have shipped 2 million copies of OS/2 2.0.
While some of those copies were bundled with IBM hardware, the vast
majority were sold directly to end users.  A conservative estimate might
be 1.5 million OS/2 users.

So, if there are four times as many Windows users as OS/2 users, has
Windows won?  Consider that there are around 60 million DOS users who
use neither Windows nor OS/2.  Compare that 60 million to the combined
total of approximately 7.5 million Windows and OS/2 users.  More than 85
percent of the potential market has yet to choose either product.
Clearly, there is a lot of ground to cover before the fat lady sings.

If both OS/2 and Windows continue at their present rate of growth, at
the end of 1993 there will be roughly 10 million Windows users and
approximately 5 million OS/2 users.  But, actually, OS/2 may grow at an
even stronger rate next year.  Here's why.

As Ziff has pointed out:

  "It's far better than DOS; it beat NT to market by about a year; it's
   fairly stable; it's backed by IBM's money; and - best of all - it's a
   great product that offers the features desktop users need today."

The reason OS/2 has failed to capture a larger market is the lack of
native OS/2 applications.  Vendors have not been developing OS/2
applications because they did not perceive a market.

Today, though, that market is growing, and those applications are being
introduced.  Lotus, Borland, Symantec, Wordperfect, Computer Associates,
and many more are introducing native OS/2 applications.  In addition, a
wide variety of excellent new OS/2-only products (Golden Compass,
FaxWorks, and Describe, for example) are also being quite successfully
marketed.  These companies are MAKING MONEY on OS/2 applications.

If OS/2 is indeed better than DOS and Windows, and has a wide range of
application vendors supporting it, then OS/2's sales growth is likely to
expand substantially.  If this comes true, then OS/2 will come close to
matching Windows market penetration in 1993.

But even if OS/2's sales do not increase substantially, it will still
have 5 million users by the end of 1993.  That's critical mass by
anyone's definition.

Again, from Ziff's column:

  "Once one standard gains critical mass, everything nonstandard tends
   to fall away rapidly (with the exception of competing standards that
   also have critical mass.)"

It is increasingly clear that OS/2 _is_ that "competing standard".

Ziff also made another point:

  "The Windows standard isn't standing still; it's evolving.....
   Windows 4.0 (code-named Cairo and due in 12 to 18 months) is the
   next stage.

  "IBM had an easy time making OS/2 run Windows 3.x applications because
   it was able to use Microsoft's own source code.  But because of its
   recent divorce from Microsoft, IBM must now write code from scratch.

  "And Windows 4.0 won't run on OS/2 until IBM can reverse-engineer it.
   That means that even with luck, Win 4.0 will not run on OS/2 until
   two or three years after it runs on NT."

As I've argued above, I expect that the OS/2 Workplace standard will be
just as important to the industry as the current Windows standard, so
one could argue that Windows compatibility will not be as important in
the future.  In actuality, though, I think that Windows compatibility
will remain quite important.

It turns out that Microsoft has made it rather easy for IBM to continue
to support the Windows API.  They have introduced a new 32-bit Windows
API for Windows/NT called Win32, and a subset of the API called Win32s
for use in the Windows/DOS environment.  Both are easily translatable to
OS/2.

IBM has contracted with Micrografx to build a new Win32 API for use with
OS/2 using extensions to their existing Mirrors technology.  This is
expected to be added to OS/2 during the second half of 1993.  In fact,
it is possible that OS/2 will support the Win32 API before any of
Microsoft's products do so.  So there need be little worry about OS/2's
continued ability to support Windows applications.

The Ziff quote above also mentions Microsoft's Cairo project.  Despite
the discussion, Cairo is not a version of Windows.  Rather, Cairo is an
object-oriented user interface that will be added to the Windows product
line.

It is unlikely that Cairo will be anywhere near ready in 12-18 months.
Microsoft has been 12 to 18 months late with every Windows version that
they've ever shipped, including Windows 1.0, Windows 2.0, Windows 3.0,
and Windows 3.1.  Windows/NT was originally scheduled for delivery
during July, 1992, and it is now scheduled for delivery during June,
1993.  Most analysts expect it to be even later than that.  So there is
little reason to expect Cairo to break such a consistent pattern.

Also, keep in mind that both the Windows and OS/2 user interfaces are
based on IBM's SAA Common User Access.  Windows is based on CUA '89,
while OS/2's Workplace Shell is based on CUA '91.  CUA '91 is also known
as the System Object Model (SOM), and has been turned over to Taligent
for use in their emerging Operating System technology as well as being
used in OS/2.

By the time Cairo becomes a part of the Windows family, the SOM-based
Workplace Shell will be running on OS/2, DOS, AIX, and the Taligent OS.
It will be running on both IBM and Apple systems, and a wide variety of
other systems containing both Intel and RISC CPUs.  Cairo will have
quite an uphill battle for acceptance.

Finally, don't overlook the importance of the actions of the rest of the
industry.  At this point in time, it would be suicidal for any vendor to
refuse to support the Windows environment.  But many of them would
certainly like to do just that.  Microsoft has alienated Windows
developers by using unscrupulous competitive techniques - like making
key technical information available to the Microsoft Applications group
long before it is available to non-Microsoft developers.  Novell's
current feud with Microsoft is likely just one of many such
confrontations.

Most application vendors would far prefer to have a "level playing
field" on which to compete, and most realize that OS/2 is likely to
provide that - since IBM is not in the applications market.

Microsoft's view of the future is one in which Microsoft Operating
System, Networking, and Application products are the only ones on our
desktops.  They intend to leverage the advantages of each of these
product areas to achieve this degree of market saturation.  End users
who are not wary of this strategy will likely find themselves unable to
easily use new products from competing vendors.

OS/2 remains a viable alternative to Windows, with a sizable and growing
market share.  Anyone who overlooks its importance is likely to find
themselves painted into Microsoft's corner without an escape.

Jim Gilliland
Manager, Tax Systems
BP America