[Table of Contents] [Search]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: "Digital Dark Age"

> 2) From their experience with mainframes Corporate Information Specialists
> were extremely aware of the costs associated with any proprietary computer
> architecture. Only one of the costs of the proprietary systems was their lack
> of a standardized storage medium. For CIS folks it was burned once twice shy.

Actually, the interesting thing is that it is almost impossible to show
that PCs are cheaper than mainframes. Mainframes (and dumb terminals)
tromp PCs on almost every single lifetime cost metric, from
installation, maintenance, support, etc. The other thing about
mainframes is that they were always backwards compatible, because they
had to be. And the are reliable. The company I consult for makes a
special purpose mainframe computer, which has an installed base of
several hundred thousand units world wide (possibly even millions) This
machine works directly with systems up to forty or fifty years (yes,
years) old, and has a average downtime measured in minutes per decade.
(Including hardware upgrades. You can do live hardware and software
upgrades on mainframes.)

CIS people fought the introduction of PCs as a standard platform more or
less tooth and nail. On some level, it was that they disliked the
reliqunishing of control, but perhaps, to be more charitable, they
understood what was being stored on computer, the business processes
that relied and produced that information. Central control is something
that businesses are increasingly coming back to, there are major
benefits to it. (not the least is much better control during the

> 3) At their heyday there were only thousands of mainframes. The PC has an
> incrediably hugh installed base today and growing fast. As I have stated over

Try millions.

> 4) Perhaps it is that in today's culture people just don't realize the
> importance of a number like 100 million.  What's another 100 million.  If an
> evil empire (Jerry Fawell-Teletubbies?) took over the United States and
> declared PC's to be illegal pornographic machines, ownership of which was
> punishable by prison, they couldn't find all the 100 million machines already
> out there in a hundred years.
> 5) How large of an installed base does it take to instill confidence in
> longterm continuity?  200 million, 500 million, a billion?  When there are a
> billion PC's all equiped with CD-ROM readers will that give you the confidence
> that in a hundred years that somebody could get one to read a funky old CD-ROM
> that their great- great -grandfather put together? I'm betting that out of
> those billion PC's a good number will remain serviceable even a hundred years
> from today.  It might be necessary to go to the antique computer store or
> vintage computer club, but the basic CD-ROM equiped computer of today will not
> suddenly disapear from the face of the earth in the next hundred years. There
> simply isn't enough landfill area for them all.

So what you are arguing here is that there is a network externality
inherent in a huge installed base. Well, sort of. The network
externality that you are talking about only described the marginal
benefit that I gain by buying a machine now. In just three years, it's
completely possible that the positive effect of the network will become
a negative effect. You also confuse what benefit the network externality
actually creates. The externality only creates benefits for new
adopters, and provides little long-term benefit. (In fact, at a certain
point, this will actually help people justify abandoning the old
technology.) The 100,000,000 computers out NOW there gives me very
little confidence that in 100 years, my grand children will be able to
show their kids the CD-ROM I put together NEXT year. Sure, they will
probably be able to read the CD-ROM, but whether they can see it as
anything other than a carefully clocked and error corrected bit stream,
I'm not sure.

Many of your arguments make an interesting defense of Microsoft's
supposed monopoly power. His argument is that the monopoly position of
Microsoft provides a significant, long-term social benefit: the ability
to ensure that the fruits of our intellectual labors remain available to
us and our descendants. These arguments specifically discount
Microsoft's claim that their monopoly position is a tenuous one, capable
at any point of being toppled by a new, innovative and different
technologies or set of technologies.

I don't think this is an easy task to solve. I don't think CD-ROMs are
the answer to the general problem of long-term archival storage of
computer data. (From an environmental perspective, they're not even a
good answer period.) The Year 2000 problem is essentially a problem in
maintaining accurate archives and ensuring that your software remains up
to date. We'll see if we've licked that problem. The bigger problem
we'll be able to report on in a few hundred more years.


Mark    -- http://www-personal.si.umich.edu/~handel
Handel  -- handel@umich.edu
        -- 805 Spring St. #2, Ann Arbor, MI 48103
        -- +1 734 669-0311

[Subject index] [Index for current month] [Table of Contents] [Search]