[Lecture Notes by Prof Yuen and possibly others 2000-2001]

From the previous lesson:

... A computer is a machine with various boxes; one of these is the
memory box that contains all the data, which it can identify by name
(actually by addresses but we wont discuss such details here); the other
is the arithmetic unit which can receive data and produce results of
various operations like +, x, etc., but you first have to pull the data
out of the memory box and send them to the arithmetic box, and send back
the results ...

History of the PC
-----------------

About 30 years ago a small company near San Jose, California, produced
a small computer processor called Intel 4004: it has an arithmetic unit 
able to combine two 4-bit numbers (i.e., 0 to 15) to produce a 4 (actually 
5) bit result. and it can pull out and send back 4-bit numbers from a
small memory box attached to it or along some communication lines that
connect to other kinds of boxes, such as a light output device that
shows a numeral 0 to 9 given four bits 0000 to 1001. Compared with the
computers being produced then by IBM, the processing power of this machine
is tiny; what makes it significant is that it was fabricated on a single
piece of silicon, rather than a whole circuit board or even a rack of
circuit boards that other computer processors must have.

Among the earliest customers of Intel 4004 were the Japanese calculator
companies: the 4004 can take two decimal numbers, two digits at a time,
combine them, and then work on the next digits; it takes 100 operations
to add two 10 digit numbers, multiplication takes longer, and division
much longer still, but since it can do a few thousand operations
per second, it is fast enough for calculators. With a few memory modules,
also fabricated on small pieces of silicon, to contain the calculator
program plus a bit of data, and 10 number displays, the whole unit can
be packaged into a small box with 20 buttons that could be sold for a
couple hundred US$. Other simple applications were quickly found, e.g.,
a 4004 can be put into a washing machine to read the dial and switch
settings, and turn on/off hot/cold water, fast/slow agitation, dirty
water pump, spin, etc. Not only was Intel making money; it could also
re-invest to make faster and bigger processors, like 8-bit instead of 4-bit,
then 16, 32 and 64 bit processors. Today the Pentium 3 can crunch several
hundred million 64-bit numbers per second, the result of 30 years of
continuing development at making electronic circuits smaller and more
reliable, and better machines to design and build more complex devices
on silicon.

The idea that you can build a small computer with such a processor quickly
arose, to make cheaper versions of the laboratory, minicomputers then avail-
able. However, merely replacing the existing processors by a smaller and
cheaper one would not produce much benefit, since the cost of a computer
lies much more in the accessories: memory, disk, terminal, operating
system, application software... The new processors only make sense as part 
of a computer in which everything has been downsized, and the system serves 
new markets not being met by the then available small computers like PDP11:
the home hobbyists. The breakthrough made by Apple Computers lies not
so much in recognizing the market, but in knowing how to do this downsizing
to meet the market.

The Applie II computer consisted of a small B/W TV set sitting on top of
a keyboard with its casing extended in the rear direction to create space
for a floppy disk unit and a circuit board holding the processor, memory
and device control components, including one that emitted a VHF signal that
went into the TV antenna line and causes letters and numberals to appear
on the screen. Except for the motherboard and the casing, every component
was a mass produced, off the shelf part that served a variety of customers.
For software, Microsoft, a small Seattle company owned by two Harvard drop-
outs, was able to supply a simple Basic compiler which had already been used
on a number of hobbyist machines, and a California company called Digital
Research supplied a system called CP/M that accepts a few simple commands
that allow files to be maintained on the disk and a particular program you 
choose to be started from the keyboard.

Around 1980, when the market was already dominated by Apple, IBM decided
to get into the act, in a light hearted way: one development lab had put
together an experimental machine; when asked at a meeting, the manager said 
it could be turned into a commercial product in a year, and was immediately
ordered to do that by his superior. A hurried effort to find an operating
system for the machine had to be made, and after failing to get negotation
even started with Digital Research (which was recommended to them by
Microsoft because of their better expertise in OS) over some non-disclosure
agreements, IBM offered generous terms to Microsoft to do the work. A
Seattle computer hobbyshop owner was found to have written a system that
runs on the processor (Intel 8088) that IBM was using, and the rights were
bought by Microsoft, which named it MS-DOS. IBM neither demanded ownership
of MS-DOS, nor patented the circuit board and packaging designs of the
PC, as it did not think the market would be lucrative. IBM also released
all technical information during the development, so that software companies
could provide programs the users might need even before the system was
released. From the very start, the ready availability of software was
an important advantage of the IBM PC.

At first, the new PC was the machine for secretaries: they could do word 
processing and email on it, and play some games in their spare time. 
One particular piece of software, Lotus 123, caused the machine to be
put on the desk of their bosses: the simple spread sheet began to be
used for project planning and presentations at meetings. Once the machines
were there, people also found it convenient to send their own emails,
access mainframes, and store files and data on them. The PC started
to cause real changes in people's work habits, and generated its own
mementum for greater use of technology and information.

Four trends already present became accelerated. First, machines had
to become much easier to use because more and more users were not
professional programmers, often unfamiliar with technology at all.
Second, the PCs had little processing capacity of their own, but
they could format a processing request to another machine according 
to that machine's requirements and get back the results for the PC
user to see, so called client-server computing. (Previously, a small
machine pretended to be part of the bigger machine to allow the user
access, but that requires a knowledge of how the bigger machine works.
Client server computing puts the responsibility on the client machine,
the PC.) Third, the spare capacity of the client machine may be devoted
to the nice packaging and presentation of the information obtained from
the server, and allows the use of multi-media presentations that deploy
sound and video facilities on the PC. Finally, the PC itself may have
just the basic software, but embedded software may be brought back
together with the data, and be used in the presentation of the information
to the user. Along with this, you could also bring back cookies, which
collect information about your activities, and viruses, which try to
wipe out your system, but these came later.

In the mean time, other computer companies realized that, since the
technical details of the PC were readily available, and nothing was
patented, it was open to anyone to make cheap copies or better versions
of the IBM PC, which must however have the MS-DOS before they could 
do anything, and so they paid Microsoft a license for every machine
sold. PC clones began to flood the market, greatly benefiting Microsoft
while leaving IBM in the cold. A fundamental power shift had ocurred.

IBM tried to recapture the PC by developing new models with different
hardware design and new operating system (PS2/OS2), but its usual
method of offering complete lines of machines attracting users with
variety, sophistication and good service into accepting high cost, no
longer worked, because the PC and earlier machines had established a
different set of user expectations for cost performance. Users stuck
to their cheap MS-DOS machines, and then put the new Microsoft Windows
system on them. Knowing that processor and memory capacities kept
growing while costs dropped, Microsoft kept pushing out software that
required bigger and bigger machines, inducing users to upgrade their
hardware, collecting a license fee for each new machine. It became
the world's most profitable company on turnovers that is only a fraction
of those of other IT companies.

The four technology trends cumulated in the web computer: you can
put all kinds of information on your computer and wait for others
to come and access it, and you can send along the software that
will run in their machines to help them look at your information.
But who want to make their information freely available and take
all that trouble to make it readable by others? Curiously, there
are many many such people, including myself. The motivation for this
generosity varies. Some do it as part of their job to educate or do
propaganda; others do it as part of a marketing strategy. For example,
your might put news and entertainment material on your machine for
free access, in order to get users to look at advertisements, which
bring you income. If your site is very popular, people might even pay
you to provide information through your machine, to get a larger
audience.

The idea of actually selling goods on the web came along soon. In USA,
people already have a tradition to buy things by mail order (people
lived in remote farms in the past), many have credit/atm cards with
electronic fund transfer features, and several delivery companies
operate world wide. You can therefore offer to sell goods on your
website, get paid by electronic fund transfer, and deliver the goods
using a courier service. Since this is so easy to do, far more business,
and far more information, are being offered on the web than there
are consumers to use them. It led inevitably to the crash of 2000.

The current situation is that we still have a flood of PCs that get bigger
and cheaper by the day, and anyone with a PC can sign up with an internet 
service provider like AOL getting onto the web, find information through 
a portal/search engine like Yahoo or Google, and do business with companies 
like Amazon. 

All four kinds of companies have valid business models, but they are all in
actually low profit margin businesses: the profit Compag makes on each home
PC is probably less than the delivery charge, while Yahoo makes $20,000
a day from the millions of pages it delivers to its users. Amazon is still
burning through hundreds of millions a year paying for huge warehouses and
distributions network, with each book sold giving a small profit. In short, 
each company can only hope to make a profit by working hard for a very large
number of widely dispersed customers, which they reach through technology. 
Despite the recent history, there is no easy money to be made in the
web business itself, only in speculation relating to them but that 
is an entirely different kind of business. 

Coming back to the PC again, you can buy a computer on the web, since
it is small enough to deliver by courier, and the pieces are standard:
a monitor (15"? 17"?), keyboard, network card, windows license... 
Michael Dell, while still an undergrad student at Austin, realized
that he could set up a company specializing in this way of selling
PCs: you keep a minimum set of components, and assemble the machine
after the order arrives and payment is confirmed, and then deliver
the machine. With low cost and high profit margin, which allowed
constant reinvestment to expand his operation to meet a rising demand,
in a few years he was a billionaire. Yet he too had to contend with
adverse conditions in 2000: others learned to compete, and his operation
depends on having a large assembly capacity so that it can meet even
peak demands quickly, which means it is not well adapted to downturns
or just expansion at a slower rate than original plan.
(See Fortune article in October 2000 issue.)

It is now said that instead of general purpose PCs, the future lies
in web appliances that do not require disk storage nor the complex
software and expensive windows licenses, and in numerous mobile units
integrated with wireless phones. PCs will return to the more traditional
role of computers, as servers that do the actual processing and data
maintenance, while the user interface is taken over by the smaller,
cheaper and more mobile devices, which depend on the information
supplier to provide the data and most of the programs needed to use
the data; dumber machines that remotely acquire their intelligence.