John, (01)
you wrote:
> Ed,
>
> I worked at IBM for 30 years, and there were many good products
> and some bad products from IBM. But there is one very important
> lesson that I learned: any technical decision that is made for
> political reasons is doomed. It may win for a while, but in the
> long run, it inevitably dies -- either by a bang or a whimper.
> (02)
Well, with few exceptions, any product, good or bad, will sooner or
later be replaced by a successor. (03)
A particular IBM technical decision, made for logistical reasons (not
political ones), effectively set the computing industry back 10-20
years, first directly, and then indirectly. Yes, the Intel 8086 (a 1972
design) was finally replaced by a 1980 hardware design in 1994, but by
that time, the dominant operating system for the IBM PC was Windows -- a
misbegotten kludge of an operating system, to fit a misbegotten kludge
of a hardware platform. Windows NT on a Pentium in 1995 was a 1980
operating system on a 1982 microcomputer platform, and Windows XP was
the first accepted approximation. (04)
Vastly better machines and operating systems were built on MC68K and
other 32-bit hardware designs with memory-mapped I/O in the mid-1980s --
early Macs, Suns, etc. But Motorola, Xylogics, and others did not have
nearly the chip fabrication capabilities of Intel, and IBM estimated
requirement for a very large number of processor chips for the PCs. So
IBM based its 1984 microcomputer product on a 1972 hardware design
instead of a 1980 hardware design, a one-generation setback from the
outset. (05)
(The claim-to-fame of the Pentium was 80-bit IEEE floating-point
arithmetic hardware, and some parallel process elements, which were
admittedly 1990s stuff, but the Pentium succeeded by 6 months the first
Intel processor with 32-bit addressability (the 486), rather than all
the 16-bit-plus-kludge stuff from the 286 and 386. It was the screwy
addressing that made it impossible to seat established and reliable
operating system designs like unix and VMS on the PC. So instead of
moving forward from VMS and Unix, we took a giant step backward with
Windows based on DOS, and hacked our way forward to a real operating
system with Windows XP. In 2002, PCs finally caught up to 1989 Mac
operating system capabilities, if you ignore the remaining kludge that
is ActiveX.) (06)
And there are lots of other examples of the success of mediocre products
over better designs, often with the result that the company with the
better design disappeared, and the design was rediscovered by a big firm
5-10 years later. It is sometimes alleged that the bigger firms also
had done work on the better designs, but had made a corporate commitment
to the inferior products. (07)
>> ... not because XML is good, but because students and
>> managers believe it is all there is. That IMO applies also to OWL. It
>> is the only modeling language W3C has, and thus it will come to dominate
>> that part of the trade that is looking for a way up from XML Schema, and
>> from the same fount of knowledge.
>>
>
> That is an argument ad hype and politics.
> (08)
Well, first, that is the basis on which many management decisions are
apparently made.
But there is another management influence at work here -- probability of
longer-term support. The better ideas of small companies tend not to
generate market until they are acquired, because no one wants to commit
some important aspect of his business to the product of a company that
could disappear after one bad management decision. In a similar way, a
standard that has hundreds of users, academic researchers and tool
builders working with it will have support, even if some of them
disappear. (09)
A theoretically better technology that is the province of only a few
small companies and high-end academics will be hard to find support
for. Can I hire people who are already familiar with it? Who is going
to teach my people how to use it? What community will provide the help
desk? What if the company that provided the tool I use goes out of
business? Is there even a commercial tool, or just university software? (010)
> Such arguments can have a major effect in killing better options,
> at least for a time. (011)
Absolutely! But it is not the arguments; it is the management practice
of heeding them. No one ever got fired for buying the big name
products, no matter how much it cost, or how much it hurt his firm's
productivity, or security. 40 million Frenchmen can indeed be wrong, but
you can easily hide among them. Most of us learned in the 1960s and
1970s that superior technology does not sell; only superior marketing,
big names, and promise of long-term support do. (012)
> But in the long run, they are overtaken by
> newer variants of the kinds of products they killed -- usually ones
> that were designed by competitors.
> (013)
Or by designs that were only one generation out of date, produced by
equally large competitors. (014)
> Bottom line: Anybody who sees that the emperor has no clothes
> has a moral obligation to say so.
> (015)
The problem is one's perspective on what constitutes clothes. Some keep
you warm and some are for show. The emperor's choice may not be
suitable for the Paris Opera, but it may meet his primary criteria. (016)
"Safe upon the solid rock, the ugly houses stand.
Come and see my shining castle, built upon the sand."
-- Edna St. Vincent Millay (017)
-Ed (018)
P.S. I agree with John in spirit, but Jared Diamond got it right -- new
technologies require the proper supporting environment to become
successful. The supporting environment for formal knowledge engineering
is just finally beginning to emerge. Let us not criticize the adoption
of the crossbow because we know about flintlocks. Be thankful that the
emperor has given up on mounted knights being the primary weapon. He
has to understand the value of firepower before he will buy what we are
selling. (019)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 (020)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (021)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (022)
|