33-1/3 to Light Speed and Back Again
My personal experience with technology
From the early years of computers
in the 1970Õs to today in 2008
By Rich Love
I was born
January, 1948, three years before the first computer (UNIVAC) existed in the United
States. It was the year that the 33-1/3 RPM record was invented. This new
record format allowed 30 minutes of music on each side of the record.
Typically, there were six songs per side for a total of 12 songs. This was a vast
improvement over the previous 45 and 78 RPM records, which only had one song per
side for a total of two songs.
The
transistor was invented just a few months before I was born.
My grandfather
was born before the flashlight was invented and before electricity was being
distributed widely as AC current in the United States. He lived in the desert
with no electricity and no running water. He drove to town once a week in his
Model A Ford to get water and supplies. During the hot summer months, he slept
in a mineshaft because it was cooler underground. Nobody had air conditioners.
My
grandfather has long since departed this Earth. I am now 60 years old.
Science
now traces modern human gnomes to fossils found in Africa 150,000 years ago.
Older human fossils have been found millions of years old.
NASA has
sent spacecraft to Mars and determined that there was water at one time. They
have also photographed Europa, a Moon of Jupiter, with a flyby and determined
that it has liquid below a frozen surface. They plan to send a submarine there
in the future to look for life in the ocean. The human genome has been mapped
using super computers, paving the way for totally understanding the
evolutionary design of the human body. None of this would be possible without
the computer.
I donÕt
think the average person appreciates this moment in time. It really is a
privilege to be part of this generation.
My
History, Experience and Observations about Computer Hardware and Software
Evolution over the past 37 years.
I attended
the University of Alaska at Fairbanks from 1971 though 1973 and graduated with
a degree in Electronics Technology. The classes I took the first year, taught
basic electronics and electronics math. The following years, I learned the
theory behind radio, television, radar and computers. In our labs, we built a
simple AM radio. We also learned how computers were designed. We built a
flip-flop (the most basic way to make a computer gate flip from 0 to 1 and back
again). The first computers were built using these flip-flop circuits. Two
transistors, and a few resistors, capacitors and diodes were used to create one
bit of computer information.
The Flip-Flop installed on a circuit
board.
As a side
note: When I entered college, I had no experience with electronics at all. My
only exposure was ignition systems in cars since I was an avid car enthusiast
and mechanic. I had two years of auto mechanics in high school and always
maintained my own cars. It was difficult for me to learn electronics because I
had previously lived in a mechanical world. Electronics seemed like an
imaginary world to me because you had to imagine electrons moving through a
wire. You had to imagine what the electrons did when they met a diode in their
path or a capacitor or resistor. I remember having dreams at night where I was
an electron traveling through a wire, blocked by a diode.
Electronic
calculators were just starting to emerge as an affordable alternative to the
slide rule in 1971. We were not allowed to use a calculator in class. That was
considered cheating. So I became very good at using the slide rule for my
electronics math calculations for designing circuits.
Slide rule I used in College
In 1972,
Hewlett Packard released the HP-35 Scientific calculator. That changed
everything. Now it was possible to do calculations quickly and accurately. It
was too expensive for me to own one, but it paved the way for less expensive
calculators that followed. I remember the absolute giddiness of the electronics
and scientific community over this calculator.
HP-35 Calculator
This time in
history was during the transition from tubes to transistors. Integrated
circuits were emerging but were so new that our electronics instructor could
only talk about them in theory. Due to equipment funding, colleges are always a
bit behind the technology curve anyway.
Learning
electronics was an important stepping-stone toward programming. Understanding
electronics and programming are very similar because you live in a virtual
world. Your thought processes for design are very slow, step-by-step actions
and the end result is something that happens instantly. It takes an incredible
amount of man-hours to design a complex circuit or program. The average person
cannot begin to comprehend the thought process that goes into designs that they
take for granted.
I learned to
write my first program on an old Athena computer that was donated to the
college. It filled a huge room and ran on many modules in sealed metal cans.
These rectangular cans were about 4 inches by 2 inches and had a handle on the
outside and plugged into a socket. That way you could replace any defective
modules. The modules contained transistors, resistors, capacitors etc. (the
same stuff we were soldering together in class to build our flip-flops, shift
registers etc). The computer had no display or printer. The only way to program
it was to push buttons on a very large console. The console was about five-feet
wide. You would enter your program as binary numbers one line at a time and
then push the run button and watch the lights blink to see if you programmed it
correctly.
Rich at the
console of an Athena computer at the University of Alaska, Fairbanks.
The next
computer I programmed was hidden in a computer room at the college. I donÕt
think I ever saw it. I used IBM punch cards to type binary instructions on each
card. Then the card stack was submitted to the computer administrator to be run
during the night. The next day I could see whether my program worked by the report
that was generated by the computer operator.
IBM Punch Card
During my
college years, I worked nights at the USO in Fairbanks. My first useful
electronic project was at the USO. I designed and installed a public address
system with a row of switches at the front desk and speakers in every room of
the USO. I had a microphone at the desk and could flip a switch to make
announcements in any room I wanted.
After
graduating in 1973, my first job was at a radar site, 27 miles north of
Fairbanks, working for Stanford Research Institute.
Physicists would also
visit the site from all over the world to perform experiments using the radar.
As I remember, the computer was about eight-feet wide and five-feet high. It did not
have a hard drive. It recorded data onto reel-to-reel tape drives. There was a
Teletype printer attached as the console device. There was no CRT screen. You
used the keyboard on the Teletype and its printed output to operate the
computer.
You could not program the computer from the console. You used a punch card machine instead. The punch card machine had a keyboard for data entry. You would program one line of computer code on each card. Then when you had all of your cards punched, you would take the stack and drop it into the card reader. Then on the console, you would tell the program to read the stack of cards and then compile and run your program. Of course, it would not work and you would have to figure out which cards had errors on them. You would fish though the cards and find the cards with the errors and re-type those cards. Take the stack back to the card reader and try again.
I wrote my first important (actually useful) program using this
method and the FORTRAN programming language. A visiting physicist, Vince
Wickwire, gave me the project. He asked me to write a program to monitor the
radar output and detect when certain changes occurred. The radar was taking
samples from the ionosphere about 300 miles above earth. Recording to tape
could go for hours. The physicists needed a way to index the tapes to look for
data. The program I wrote tracked changes in the radar azimuth and elevation.
It also looked for significant changes in electron densities and ion
temperatures. This was a great learning experience for me. I did not know the
first thing about FORTRAN, but Vince taught me enough to get started. I figured
out the rest on my own.
The second electronic
project I designed was a flash adaptor for the Polaroid SX-70 camera. I think
this was in 1972. Polaroid cameras were popular at the time but used flash
bulbs instead of an electronic flash. You would buy a flash bar that had five
flash bulbs. After five pictures, you would buy another flash bar. That was
expensive. I designed a circuit that would plug into the flash bar socket. It
took the output from the Polaroid and connected it to a standard electronic
flash socket using a diode, capacitor and a transistor. I canÕt seem to find the
schematic I drew for this design, but I still have the flash adaptor. I used
the flash bar body but removed all of the flash bulbs and installed my
circuitry inside. Then mounted the electronic flash shoe on top of the flash
bar.
SX-70 Polaroid Camera
RichÕs Polaroid SX-70 Flash Adaptor Prototype
In 1975, I packed up the
family and we moved to Washington State.
1976, my next job was
with CX, a film processing company. They made equipment for the photo processing
industry (barcode readers, printers and terminals).
This was before digital
cameras existed. Film was still a big business. Film processing plants would
track orders by using a film bag with a barcode at the bottom of the bag. The
unprocessed roll of film would be placed in the bag at the retail store when
the customer dropped off the film to be processed. When the bag arrived at the
film processing plant, an operator would insert the bag into a CX-7020
printer/scanner. The barcode was read and a price was printed on the bag. The
transaction was recorded in the computer. A terminal with a single line display
would display some information about the order for the operator.
At the CX manufacturing
plant, I would test the entire system before it was shipped to the processing
plant. The PDP-8 (and later PDP-11) computer was purchased from DEC. I would
test the computer with all of the peripheral equipment connected to it before
shipping it to the processing plant. As I mentioned earlier, the peripheral
equipment included the CX-7020 printer/scanner and a terminal. The printer and
terminal were manufactured at CX by low-wage assemblers and delivered to my
final test department. The circuit boards were all assembled and soldered by
hand. It was absolutely guaranteed that the printer and terminal would not
work. There were just too many possibilities for errors. The circuit boards
would have solder bridges between pins, IC chips and other components would be
bad, wires would have bad connectors or be wired incorrectly. I used an
oscilloscope and schematics to troubleshoot and repair the printers and
terminals. I became so good at this that I later wrote the service manuals for
all of the equipment that I repaired during my four years in the test
department. I wrote nine manuals total.
The field service
technicians and the final test department used these manuals. I think it was
1981 when CX got their first circuit board testing equipment. It was a
programmable machine that you could program to test circuit boards. You would
lay the board down on a flexible bed that had a vacuum to lower the board onto
a bed of nails. Each one of the nails in the bed would contact the correct spot
on the underside of the circuit board to test it.
This created a whole new
department of technicians. Their job was to take the printout from the testing
machine and replace the suggested component and/or repair solder bridges. Now a
reliable circuit board could be installed in the printer or terminal before it
was delivered to final test, making their job much easier. This was the
beginning of the end for electronic technicians with my skills. Now computers
were starting to analyze computers to let the technicians know what to repair
or replace.
Rich testing a CX 7020 Barcode printer/scanner
While at CX, I learned
to program the PDP-8 computer in Pal-8 assembly language. I started by learning
how the existing test programs worked and then modified them to add more
features. Then I wrote a few of my own test programs. PDP-8Õs were manufactured
from 1965 through 1974.
These computers had a
panel of switches that you used to boot the computer. You would flip the
switches to set the correct address and store or read the data in that address
position. You could actually write an entire program and run it from those
switches. However, it was more efficient to use punched paper tape to enter
your program. You could also type in the program using a Teletype. It was not
until the late 1970s that CRT terminals were in widespread use with the
release of the VT50 CRT terminal.
The computer did not
have a CPU chip. It used a set of three logic boards to create the CPU. The CPU
could only address 4K of memory at a time. The CPU handled fields of 4K
memory. A program would typically
take 3K of memory, leaving 1K for user space. But you could have more than one
program running (one in each 4K memory field) up to eight fields. This was the
beginning of a multi-user system because you could assign one user to each
field.
PDP-8 Control Panel
The PDP-8 used core
memory. This was very expensive memory consisting of tiny ferrite iron cores
that looked like little doughnuts. Tiny wires passed through the core to
magnetize it in a positive or negative direction. Then the memory was read by a
sense wire that would detect the polarity of the charge and determine if the
value was a one or a zero. This was one bit of information. The circuit board
containing these cores was 10.5 inches by 9 inches. Each core represented one
bit of data. The cores were installed on stacks. Each stack had an array of 64
x 64 cores for a total of 4096 cores. But to get a 12-bit word, you needed 12
stacks of cores. These stacks were sandwiched on top of each other on one
circuit board to produce a total of 49,152 cores. Divide 49,152 by 12 and you get 4096 12-bit words (or 4K). The whole 4K board could store
4,096 characters . Each character required 12 bits of data or 12
cores. Eight bits were used for the ASCII code of data and the remaining four bits
were used for error checking and other functions. You could daisy-chain up to eight
boards for a maximum of 32K. Today, it is very hard to imagine a computer that
could run on 32K of memory. But these computers cost $18,000 with 4K of memory
and no peripherals.
Very early core memory design pre-dating the PDP
computers.
To read a character, the
computer would activate the sense lines going though 12 ferrite cores and
detect the zero or one value. Using the binary numbering system, you could
store and read characters. For instance, a character is represented by reading
the first eight core values (the last four cores were not used for the character, but
were used for operation codes). DEC used octal math to describe characters.
These were in groups of three. However, I have always thought of them in groups
of four, which is BCD or binary coded decimal, which I mentally convert to
decimal.
Here is the letter 'A':
0110 0101
That is a decimal 65,
which is the ASCII code for the letter 'A'.
Using binary arithmetic,
you can decode the zeros and ones in the binary number. ItÕs really not hard to
understand (well, maybe). The numbers are in groups of four.
Computers still use this
same logic to read and write characters.
Core memory boards were
originally used in the ENIAC computer in 1949. They saw their last use in the
late 1970s. This is an amazing technology to me. These memory boards were
originally handmade using a stereo-microscope. The ferrite cores were set in
place and tiny wires were threaded through each core.
Since memory was so
expensive, the programmers had to be very clever about memory usage. Virtual
memory was used on the disk drives to have enough room for running software. It
is hard to imagine a computer today running on 4K of memory. But in the
1970s, thatÕs what they did. The early PDP-8 computers used punched paper tape
to read a program. Or you could type in a program using a 110-baud Teletype.
There were no CRTs connected to computers yet.
While working at CX, I
taught myself the BASIC programming language and used it to program a
manufacturing job control system from scratch. I also wrote a Bulletin Board
System in my spare time. I wanted it to be an in-house system for the employees
to use, but never got approval for its use. Bulletin Board Software was used at
that time (before the invention of the internet). You would use a dumb terminal
and a modem to dial into BBS systems and leave messages and respond to others.
In 1977, I purchased a
used terminal and set it up in my kitchen at home. I built my own 300-baud
modem using schematics I found somewhere. At first, the only thing you could use
it for was to dial into BBS systems. But later, Compuserve came into existence
and that became more popular than BBS systems. America Online was written for
Macintosh in 1989 and I was one of the first members and still am, just for old
time sake.
In 1978 I purchased my
first computer. It was an Apple II.
The Apple II had 4K of
random access memory expandable to 48K.
First Apple I Computer Designed by Steve Wozniac
Apple II
In 1982, I went to work
for Microdata as a Field Service Engineer. I drove all around the Seattle area
repairing mini-computers and peripheral equipment at customer sites. Companies
were still paying over $100,000 for a computer in those days. The personal
computer was just starting to emerge but was not in wide use in business yet.
Most businesses had a mini-computer with dumb terminals connected to it via
RS-232 serial cable. The main printer was attached to the computer. It was
quite often a Printronix 300-line printer. This was a 300-lines-per-minute dot
matrix printer. Each user had a terminal and sometimes had a small slave
printer attached to the terminal. The terminals were called dumb terminals
because they were not a computer (they do have a program running in firmware
but are not programmable and have no storage). They had a keyboard and a CRT
display. When you typed on the keyboard, you were actually sending each
keystroke to the mini-computer via the serial cable. The mini-computer would
echo the character you typed back to your CRT screen.
The mini-computer would
also send escape and control characters to the CRT that were invisible to the
operator. Those sequences would tell the cursor where to go on the screen,
change screen colors and character colors, erase characters or whole blocks of
characters on the screen. It could designate areas of the screen that were
protected and then do an entire screen erase leaving the protected characters
on the screen. This was purely a text-based system with no graphics. The
closest thing to graphics was that it would place special drawing characters on
the screen to form boxes around text. These systems actually worked quite well
and there are many businesses using dumb terminals today in 2008.
The mini-computers stood
about five-feet tall and were about two-feet wide and three-feet deep. This was a rack
system. With the covers on, you could not tell that there was a rack inside.
But when you took the covers off, you could see that the computer only took up
about one foot in height of the five-foot tall system. It slid into the rack on rails.
You could pull the whole computer assembly outward to get at the circuit
boards.
Each board was
approximately eight-inches wide. There was a CPU board, memory board, serial
interface board, tape interface board and disk interface board.
Mini-Computer Rack
Actual size, five-feet high
The tape drives needed
constant maintenance. You had to clean the read-head with an alcohol pad at
least once a week. The stepper motors used to drive the reels would require
electronic adjustment using an oscilloscope every few months. The alignment of
the tape to the head would also need adjustments. When the drive went bad, I
would figure out which part was bad and replace it. Then re-calibrate the drive
with the oscilloscope again. There were many replaceable parts: read/write
head, reel motors, electronic circuit boards, servo motors, tension arms,
rollers etc.
Microdata tape drive
The disk drives were
huge in those days. Internal disk drives that could mount inside of the main
rack were about two-feet wide and 2.5-feet deep and six-inches high. They weighed
about 80 pounds and consisted of large multiple platters that were about one foot
in diameter. There were several circuit boards in the unit to control the
drive. A typical drive like that would hold 10 MB of data. During the next few
years, the drives stayed the same size but the capacity increased to 50 MB of
data. Compare that to drives in 2008 that hold eight GigaBytes of data on a USB
thumb drive that you put in your pocket or on your keychain and cost $50. ThatÕs 160
times more data than the 80-pound drive of the 1970s costing $10,000.
Microdata Reflex I drive. 50 Megabytes, 80
pounds, $$$
Those disk drives were
very expensive and were expensive to maintain. When the drive crashed, due to
the floating head coming in contact with the platter, my job would be to open
the drive up and replace a platter and the head. Then use an oscilloscope to
re-align the head to the spinning platter.
Larger external drives
were affectionately called washtub drives. They stood about four-feet tall and
were about two-feet square. They had a removable cartridge full of platters.
These cartridges typically had a capacity of 50 MB. A large computer room could
typically be 50-feet square with washtub drives and tape drive racks all over
the room.
Today in 2008, you can
get a solid-state USB flash drive that is one inch long, holds 8000 MB of data
and is portable (put it in your pocket). Hard drives are five-inches long and hold
750 GB of data. When a drive goes bad, you throw it away and buy a new one.
As a field service
engineer, the knowledge and skill required to repair computers in the 1970s and
1980s was at a much higher level than todayÕs computer technicians. You were
actually a combination of technician and engineer. The equipment was so
expensive, it was worth the time and money to repair. Customers would typically
pay $5000 per month for a maintenance contract that covered their computer,
terminals, printers, tape drives and disk drives.
While working for
McDonnell Douglas as a Field Engineer in the Seattle area, I was sent to
southern California many times to learn how to maintain and repair new
equipment as it was released. Whenever a new disk drive hit the market, I would
be back in California for several weeks to learn it. The classes were very
in-depth. The instructor would go over each schematic for the circuit boards
and explain the complete logic of the circuits. We would use oscilloscopes to
troubleshoot the boards and make adjustments. There was a lot of lab time where
the instructor would plant a bad part in the drive or misadjust something to
break the drive. Then I would figure out what he did and fix it by replacing
the bad board I found or making an adjustment. The disk platters and heads were
replaceable items also. The drives usually cost about $10,000.
The most common business
printer was the Printronix 300-line printer. It was a dot matrix printer that
used 132 hammers with little points at the end of each hammer. When the hammer
would strike the paper, it would make a dot.
There were several
circuit boards that controlled the shuttle, hammers and paper feeding. The
hammers would get weak or break. They were all individually replaceable. The
shuttle was necessary to print the width of a character. If a character was six
dots wide, the shuttle would move over by the width of six dots and then move
back again. During this horizontal movement, the hammers would fire in the
correct location to form a character. A common problem with these printers was
misaligned characters. They would get wavy looking when it was not aligned
properly. Using an oscilloscope, I would adjust potentiometers on the circuit
boards to do the alignment.
Printronix Printer with Sequel mini-computer and
Terminal
Programming
As I mentioned earlier,
the first programming language I learned was FORTRAN while at Stanford Research
Institute in Alaska during the early 1970s, using IBM punch cards. I wrote the
LIST-A-CHON program for cataloging tapes recorded by the radar system.
In the late 1970s, I
was working for CX Corporation as a test technician for photo processing
equipment. During that time, I taught myself assembly language on a DEC PDP-8
mini-computer. Programs that tested the functionality of the photo processing
equipment were written is assembly language. I modified these programs to
improve the tests. When equipment design changes were implemented, I changed
the programming to work with the new designs.
I also taught myself
BASIC, using the PDP-8 computer. I programmed a Bulletin Board System in BASIC
on a Data General computer at CX and wrote a biorhythm program in my spare time.
In 1978, I bought my
first personal computer, an Apple II. I converted my biorhythm program to Apple
II BASIC.
My first commercial
program was a lotto-number-picking program for the Apple II called Likely
Lotto. I sold it at Empire Electronics in Burien, WA. I loved that store. In
the early 80s, it was the only Apple Store in the Seattle area. It was such a
thrill to see new products come on the market for the Apple II. It was a newly
emerging market and there was not that much available yet. It made every
product release special (mice, joysticks, disk drives, monitors, software).
My Likely Lotto software
was on a 5-¼ floppy disk. My wife, Karen, designed the package cover and
I made the disk copies myself. I later converted it to run on Apple IIGS on a
3-½ floppy disk.
3.5-inch Floppy Disk
Terminal Emulation
Software
In 1986, I purchased my
first Macintosh and learned to use Microsoft QuickBASIC. A few years later, I
was working for Microdata as a Field Service Engineer. I drove all over the Seattle
area repairing mini-computers and peripherals. A couple of my co-workers were
working on a Prism terminal emulation program for their Atari. I wrote my own
version for Macintosh and called it MacPrism. I soon discovered that someone
else was using the name MacPrism and changed the name to MacToPic. The
Microdata systems ran on PICK software, but Microdata had their own version
called Reality. MacToPic was a play on words. I was afraid to use the word
'PICK' in the name because Dick Pick was known for suing anyone who did. The
PICK system was a relational database system and quite popular with businesses at
the time.
In 1990, a consultant
for Bank of America heard that I had written MacToPic and asked me if I would
modify the program to transfer data to and from the bank system that was
running on a Microdata computer. They also needed Prism terminal emulation.
Bank of America banks in the Seattle area were using Mac SE computers with a
little nine-inch screen. MacToPic had to work on those small screens as well as
the Mac II 12-inch screens.
I wrote the data
transfer routines to convert the PICK data format to a Macintosh format that
you could save as a Microsoft Excel spreadsheet.
After that success, I
realized that there was a market for my software in the PICK business
community. I obtained a mailing list and started promoting MacToPic.
In 1995, companies were
not willing to pay high prices for software anymore.
A few years later, my
MacToPic sales were so low that I decided to include most of the features in
MacWise and retain the $95 price.
Today, in 2008, I am
still selling MacWise (18 years of terminal emulation programming and sales)
and sales are increasing due to the increasing market share of Macintosh
computers.
I have gone though some
major conversion programming during that time.
All of this work has
been worth the effort.
Back to 33-1/3
Records
It turns out that there
is something very appealing about vinyl records. I recently took my old record
collection out of the closet and set up my old record player.
My Hope for the
Future
Computers have been very useful, enabling scientists to discover who we are, what we are made of, where we came from and where we fit into the cosmic universe and beyond. Not all of those questions have been answered yet.
It is my hope that technology will be used with intelligence and compasion to make the earth a better place to live for everyone.
Religious fanaticism is
a great danger on Earth. Wars continue to kill people needlesly, justified by religions based on superstitious
legends and who has the best invisible friends. To me that is completely
absurd.
Use technology to educate
ourselves and ensure the survival of the earth and the human race.
Use technology to find new sources for energy and stop the danger of wars and financial crises caused by our limited oil resources.
Use technology to make
the earth a greener, safer, a more enjoyable place to live.
Use technology to prevent disasters that can extinguish life on Earth Š asteroids, comets, earthquakes, water shortages and food shortages.
Find the cure for cancer, heart disease and other life-threatening illnesses.
It seems to me that we
have plenty to keep us busy surviving against the forces of nature in the
universe, without fighting each other.
I have the greatest
respect for scientists who devote their lives to searching for the truth,
regardless of political and religious influences. Making life more enjoyable for
the short time we have to spend on Earth and ensure the survival of our
offspring Š that is the best use of technology.
I know that this is
naive, wishful thinking, but it is something we should strive for nonetheless.
Rich
Love
Carnation Software
|
