Here, Crockford takes us on a tour through the history of computer input devices and programming to understand how we got to the current state-of-the-art.

Tramiel was fond of saying “business is war”, a sentiment no doubt owed to the trauma of growing up in a Jewish household in Poland in the early 20th century. During the German occupation, his family was relocated to the Jewish ghetto and then to the Auschwitz concentration camp, where he was examined by the infamous Dr. Mengele. After the war, Tramiel moved to the US & joined the military, which gave him the means to start a typewriter repair business and then get into typewriter manufacturing. From that simple beginning, Commodore Business Machines went on to produce adding machines, watches, digital calculators, and finally computers.

Jack Tramiel (courtesy Wikipedia)

Like a good military leader, Jack surrounded himself with innovative and hard-working people and got out of their way…for the most part. Bil Herd tells a story of Jack poking his head into the lab where he was building the Plus/4. “How many chips are in there?” “Eight”. “Make it five.” And then he walked out. He didn’t obsess over the technology. Tramiel wanted products people would love, but not so cutting edge they were impossible to afford.

“We need to build computers for the masses, not the classes.” — Jack Tramiel

In addition to outstanding talent, Commodore owed much of their success to Tramiel’s belief in vertical integration, most notably the purchase of MOS Technologies in 1976. MOS Technologies was the creator and manufacturer of the affordable 6502 CPU, which cost only $25 in 1975. For comparison, Intel’s 8088 was priced $100 higher. Owning the primary manufacturer of 6502s and having the ability to produce completely custom microchips on a moment’s notice let Commodore underprice and outperform the competition.

Jack Tramiel certainly had a dark side. If business was a war, Commodore had to win it. No trick was too dirty, and suppliers would find themselves taking a loss on Jack’s cancelled purchases, only to end up selling them to Jack’s minions in the clearance market. He was also famous for “Jack Attacks”, where underperforming employees were fired on the spot, sometimes en masse.

When Commodore’s Chairman, Irving Gould, fired Tramiel as CEO of his own company, he turned around and bought their rival Atari. With Commodore’s Amiga and Apple’s Macintosh ushering in the 16-bit era of home computing, Tramiel put the company’s weight behind the Atari ST line of computers. These black-and-white computers shipped with TOS (“The Operating System”, or “Tramiel’s Operating System” informally). The “Jackintosh” never caught the mainstream’s interest, but the ST line’s built-in MIDI interfaces made it the darling of early electronic musicians.

In retirement, Jack and his wife Helen (also a concentration camp survivor) enjoyed a life of luxury and philanthropy, funding Holocaust-related museums and causes. Some people loved him, some people hated him, but you can’t deny the effect Jack Tramiel and Commodore had on the early home computing market.

Though he didn’t have much involvement with the computing field in his later years, Tramiel did appear at a Computer History Museum event celebrating the 25th anniversary of the Commodore 64.

For more about Jack:

CNet’s interview with Tramiel in 2007

Jack Tramiel — Survival and Starting Over

Since Steve Jobs passed away last year, there’s been a lot of books put out about the man and his amazing company. But this one promised to be different. Bartending isn’t about closed-door meetings and Jony Ive’s secret studio. It’s a collection of stories from a former Apple Store Genius, the manager of his particular Genius Bar in fact. Anyone who’s worked a retail job, whether in the computer industry or not, can really appreciate the anecdotes in this book. It’s a side of Apple most people never hear about.

Bartending: Memoirs of an Apple Genius

Despite its unique premise, I was disappointed by the finished product. Coming in at just 46 pages, the book can be read in under an hour, and these memoirs are just anecdotes, often just 2–3 pages long, about funny things that happened on the job. The individual stories are delightful, and give a real sense of the kinds of issues the public face of Apple’s support department face every day, from college students mistaking MacBooks for urinals to Apple’s poorly-designed G5 towers. But the author could have included more stories, or at least embellished the ones he included, to flesh out the book as well as the reader’s understanding of what really happens in those clean & brightly-lit stores full of smiling staff. It’s no wonder this book was released early — what kind of last-minute editing was needed to get this out the door?

Still, this self-published work is only $8.99 and definitely gave me some good laughs tonight, so I still recommend picking it up and enjoying it for what it is. But it could be so much better, and I wish this book much success so Hackett is encouraged to follow it up with the kind of book this subject deservers, by expanding on his own memories and letting other former Geniuses contribute their own stories.

Bartending is available for the Kindle at Amazon.com.

It’s also available as an ePub (readable in iBooks) from the author.

IBM’s 1401 was one of the first computers to be built entirely using transistors instead of vacuum tubes. The new solid state technology made the 1401 smaller and more affordable than its predecessors. It was also much easier to program. Those attributes made it incredibly attractive to businesses in the early 1960s.

Read IBM’s tribute to the iconic 1401 family

The machine’s popularity was a blessing and a curse. Developers loved working with the 1401 and its peripherals, especially the 1403 printer. Corporate leaders and bureaucrats wanted to see what the futuristic new machine was could do, owing to both professional due diligence and human curiousity. This combination of programmers eager to show off and stakeholders hungry for a demonstration they could understand led to the birth of a nacent demoscene.

IBM 1401 Family

It wasn’t long before programmers noticed a “cellolike hum” when an AM radio was brought close to the 1401’s memory, and the 1403’s new “chain printer” technology was capable of belting out mechanical reditions of classical tunes and standards. These early digital artistic endeavours still captive audiences today:

That […] is why IBM staff, who had come to regard the 1401 as a musical friend over the years, held a “funeral” for the computer when it was decommissioned.

“I think the engineers formed an attachment to this machine above others they worked with precisely because they gave it these ‘human’ qualities — the ability to make music,” he said. “They felt they couldn’t just throw it away […]”

Music wasn’t the only artistic outlet for 1401 programmers. The 1403 printer offered a new dimension of control over the computer’s output. And in the hands of less academic users than previous machines, it was put to rather prurient use:

[…] some clever operators figured out that it was possible to “do rudimentary ‘graphics’ using only the characters available on the print chain and different degrees of overstriking to get darker and lighter areas.” This graphics technique was mainly utilised to print pictures of Playboy models, Mona Lisa and the moon (this being the Space Age).

- Humans Thinking Like Machines (emphasis mine)

EDITH was another popular graphical program for the 1401. Some sources claim, but I can’t confirm, that EDITH was available from IBM itself. Regardless, this program was immensly popular and traded openly among programmers, who still remember it fondly.

IBM 1401 sense switches

The EDITH program relied on the “sense switches” on the 1401’s operator’s panel:

It would print a picture of a woman in evening dress accompanied with the text: THIS IS EDITH, ANOTHER OPTIONAL FEATURE OF YOUR IBM 1401. IF YOU WOULD LIKE TO LEARN MORE ABOUT HER RE-RUN THE DECK WITH SS A. The program then proceeded through selections made by flipping the printer “sense switches” A, B, C, D and E, controlling EDITH’s different states of clothing. The B switch would produce EDITH in skimpier clothing with a top and short-skirt, followed in C by a tiny bikini. If the operator should be so daring as to proceed to D after the warning, “WARNING: FURTHER SWITCHING OF SS D IS NOT RECOMMENDED!” he would be treated with the anti-climax of an EDITH holding up a “modesty” sign saying “SORRY, YOU CAN’T DO EVERYTHING WITH A 1401. (NO MATTER WHAT OUR SALES FORCE MIGHT SAY.)” Although rumours tell of how running the program with the E switch would actually show a totally nude EDITH with a caption like “WELL, MAYBE YOU CAN DO ANYTHING WITH A 1401.”

- Humans Thinking Like Machines

For more information on the IBM 1401, please see the tribute site and blog from the Computer History Museums 1401 restoration project at ibm-1401.info.

The Computer History Museum also held a tribute to the 1401 in honor of its 50th anniversary.

And no, I don’t have any pictures of EDITH to share. Sorry!

These engineers took their inspiration from Star Trek and other science fiction stories, where people of all walks of life could converse with a computer like another member of the crew. Or, in the 1968 episode The Ultimate Computer, Captain Kirk actually talked a computer to death. Creating this kind of experience with 20th century technology meant making major strides on two fronts:  natural language processing (NLP) and speech recognition.

Natural Language Processing

ELIZA was a notable early experiment in NLP, built between 1964 and 1966 in the infancy of interactive computing. Though ELIZA is synonymous with a simulated psychiatrist, her Rogerian dialog was driven by the DOCTOR script, one of the many personalities which could be loaded. (UPDATE: For more on Eliza’s history, implementation, and reception, see The Digital Antiquarian’s three-part blog series: part 1 | part 2 | part 3)

A session with ELIZA

Shortly after ELIZA’s debut, Terry Winograd’s SHRDLU (1968–1970) let a person control an artificial intelligence as it interacted with a virtual world. It understood commands like “FIND A BLOCK WHICH IS TALLER THAN THE ONE YOU ARE HOLDING AND PUT IT INTO THE BOX.”, and it could respond to questions like “ HOW MANY BLOCKS ARE NOT IN THE BOX?”.

After these early experiments, NLP research shifted towards conversational expert systems. Recognizing that general-purpose AIs were far beyond the reach of 1970s technology, most researchers instead pursued AIs geared to specific subjects or tasks. For example, an expert system might monitor stock levels in a warehouse and tell the manager when to order more widgets. But, it wouldn’t be able to tell you if you need an umbrella tomorrow. Such an AI was still a long way off, pursued by only a few dedicated scientists who were able to get funding.

While research continued in state-of-the-art labs around the world, the computers landing on individuals’ desks in the 1980s paled compared to what was available to researchers, both in terms of storage and processing power. Personal computer enthusiasts rediscovered ELIZA, as its primitive action-reaction logic was well-suited to 8-bit processors and minimal RAM. But the idea of having a conversation with your PC was still a ways off.

Speech Recognition

The Lipstik was so novel, they advertised it on the box as one of the game’s features. Some say it was the only fun part of the game.

While hardware resources meant NLP languished on the desktop, consumer speech recognition received a boost from the wealth of expansion ports on those early 8-bit micros.

Before diving into the practical side of speech recognition, let me tip my hat to Access software. Their 1987 spaceflight simulator Echelon shipped with a novel controller called The Lipstik. This lightweight headset enabled wannabe pilots to launch a barrage of missiles at virtual enemies by shouting “FIRE FIRE FIRE FIRE FIRE”.

Or, alternately, “SHOOT SHOOT SHOOT SHOOT SHOOT”.

Or, “NO, MOM, I’M NOT USING MY INSIDE VOICE.”

At its heart, Lipstick was a simple circuit that measured amplitude and sent a signal when it crossed a certain threshold. Any loud sound triggered a missile, whether it was a person speaking or a dog barking. It was incredibly primitive, but it added a whole new dimension to gameplay, albeit a loud and sometimes annoying one.

Two years earlier, the Covox Voice Master debuted for Apple II and Commodore microcomputers, followed shortly by a version for Atari’s line of 8-bit micros. Retailing at $89.95 (just shy of $200 in today’s dollars according to DollarTimes.com), the Voice Master included dedicated hardware for digitizing audio and playing it back with much higher quality than the computers themselves. It also included a demo disk with software for basic speech recognition. It could pick out single words, but was still constrained by what home computers could do.

My wife and I get a lot of mileage out of this quote from The Simpsons:

1969: Man walks on the moon. 1971: Man walks on the moon… again. Then, for a long time, nothing happened. Until tonight.

It fits so many situations, especially the state of desktop computing from about 1985–1991. Don’t get me wrong, there were major leaps during this time: 16-bit and then 32-bit processors went mainstream. Kilobytes of RAM gave way to megabytes. Digital sound went 16-bit, and graphics made huge leaps in resolution and color depth. But we mostly just saw refinements of the same desktop computing experience. The early Macintosh soon had competition from more capable GUIs by Amiga and Atari. Windows grew from a novelty to the very capable Windows 3.1. We still had windows, icons, menus, and pointers, but they were better.

At the dawn of the 1990s, Intel’s 16MHz 80286 processor was a popular choice for system builders. By its close, 500MHz Pentium IIIs were pretty common. In 1994, researchers at SRI International (a DARPA contractor) left to form Nuance Communications, with the goal of commercializing speech recognition technology they built for the US military. Commercial computing technology had caught up, and they brought speech recognition to the masses.

Enter Siri

Still, speech recognition was largely a novelty for years, used by people who physically couldn’t type and those who just wanted to say “look, Ma, no hands!”. That changed in 2011 when Apple debuted Siri as part of the iPhone 4S.

Interestingly, Siri is the brainchild of another group of SRI alumni who wanted to commercialize the Cognitive Agent that Learns and Organizes project they built for the military. Siri, Inc. leveraged Nuance’s speech recognition platform to deliver a personal assistant on Apple’s smartphones, and Apple saw their app’s potential. In April, 2010 they bought Siri, Inc. and integrated their personal assistant software into iOS.

Siri’s success is reflected in the number of people hacking around limitations that prevent the software from running on Apple’s older smartphones, and the numerous clones which have sprung up for iOS and other platforms. But, Siri’s blend of state-of-the-art SRI research in NLP and speech recognition is providing millions of people with their first taste of voice-controlled computing and its promise for the future.

Since I feel guilty for taking a whole month to deliver this post, here’s a bonus video showing early speech recognition being used in the treatment of an autistic person in 1984. Hopefully Siri will inspire further research, and the technology we use 30 years from now will seem as basic as what we see here.

Which will probably delay the Covox/Siri thing even longer. Sorry.

And just what is “edgy”?

Just imagine, there used to be rooms of trained typists, secretaries…call them what you will. They typed up handwritten notes and recorded dictation with all the right formatting and few, if any, errors. These typing pools seem so quaint now, but back then typing was something you went to school for and even got certified in! The flip side of this, though, was that few other people in the business world knew how to type themselves. And they didn’t see the point in owning something that prominent featured a keyboard and came with sofware to do word processing. They wanted something they could control as easily as their TV set, or even talk to.

[…]it’s worth noting that back in 1979 people viewed the keyboard as an impediment to using computers. After all, only secretaries could type and the rest of us need to be able to talk to the computer. Hence the decades spent on trying to get computers to understand speech.

- Bob Frankston, Implementing VisiCalc

VisiCalc is regarded as the first “killer app”, replacing days’ worth of manual calculations with what we now know as an electronic “spreadsheet”. Lotus 1–2-3 built more business-oriented functionality into the same basic interface. By working with numbers instead of words, math instead of grammar, these applications helped land early Apple IIs and IBM PCs on employee’s desks. Some even brought their own computers from home in order to take advantage of these time-saving miracles.

Early Macintosh sales brochures

Mice and Graphical User Interfaces furthered computers’ acceptance into offices around the world by taking the focus away from the keyboard. Early Macintosh brochures rarely showed the keyboard on the cover, focussing instead on its friendly “face” and its mouse. But Microsoft and others soon brough such ease-of-use to x86 PCs, and both camps invested heavily in suites of office software, empowering executives to create documents on their own.

So where’s the relevance to modern computing? Join me next week for Covox, Siri, and Beyond. Until then, this promotional video from Microsoft demonstrates what happens when executives have the capability of making their own business documents. This really happened. Everywhere. It was kinda creepy.

Note: This is actually a much later video from the 1990s, supplementing Sculley’s original vision with a framing presenter and more use cases for the future technology.

While the overall idea was the domain of science fiction in 1987, the individual pieces needed to make it a reality were well into development. Knowledge Navigator was, in fact, a conservative bet to someone immersed in the high technology of the age.

Apple’s new Siri product, born of research from the prestigious SRI, showcases how far we’re come in 25 years. There may not be a man in a bow tie personifying our computer, but we can now interact with our information conversationally, and that’s powerful. For days following Tim Cook’s Siri presentation, bloggers were quick to call Siri the embodiment of Knowledge Navigators vision. What’s really spooky is, as Andy Baio pointed out, there are things in the Knowledge Navigator video that point to it taking place in September 16, 2011 — just shy of Apple’s October 4th announcement.

This isn’t the first time comparisons were drawn between Knowledge Navigator and the present-day state of the art, however. We all want to believe we live in the promised future. Jim Carlton’s 1997 book Apple proclaims:

A decade later, the Internet’s World Wide Web would explode, providing the same ability to scan vast databases as the Navigator promised while communicating with other people through the computer at the same time. […] Sculley was also years ahead of his time in predicting the importance of big 1990s trends such as the computer’s new ability to combine video and sound. For all his later failing as CEO, Sculley would go down in computer history as a great visionary and seer.

It seems Knowledge Navigator is still serving it’s original purpose. It’s providing us with a vision of the future and giving us something to measure our progress against. It may be another 25 years before we see something exactly like it, but developers will keep striving toward it and continue celebrating each part they successfully create.

Since it’s a holiday weekend, here’s some additional reading/viewing that’s well worth your time:

Knowledge Navigator Implications (1988) : Steve Wozniak, Alan Kay, Ray Bradbury, Alvin Toffler and others pontificate on the features included in the Knowledge Navigator vision and how we might interact with computers in the future.

Bud Colligan weighs in : In the comment section of Andy Baio’s piece, Apple’s Director of Higher Education Marketing from 1985 – 1988 describes the genesis of the Knowledge Navigator concept.

The Making of Knowledge Navigator : The video’s co-creator describes the ideas & some details behind how the video was made.

Dave Greelish interviews John Sculley (part 1) (part 2) : A late 2011 interview with John Sculley, revisiting his days at Apple and the decisions he made. They also discuss Alan Kay, his Dynabook concept, and the Knowledge Navigator.

So let me sum it up: a Jackson Pollock painting is Jimi Hendrix at Woodstock.

Jimi Hendrix (1942–1970) grew up in a world where recorded music was everywhere, and new technologies brought near-perfect “high fidelity” stereophonic recordings into the home. Not only could you buy your favorite song, you could buy your favorite performance of a song, and musicians had an incentive to make each performance different from the studio version to sell more tickets & “live” records.

Freed from the need to make each performance the same as the last, musicians were able to reduce a performance down to its essentials. Instead of sounds organized by easily performed measures, refrains, and codas, musicians like Hendrix used their instruments as amplifiers of emotion, props in a audio-visual display of raw feelings. Rock out a soulful solo, play the guitar with your teeth, then set it on fire.

Jackson Pollack (1912–1956) died while Hendrix was still a kid, but he came of age in the heyday of the Brownie camera, a simple camera whose $1 price tag made it ubiquitous. His drip period started about ten years after color film hit the market. To that point, most paintings were portraits of people or moments frozen in time. Even pointillists drew recognizable forms. But affordable color cameras meant anyone could capture their own images of Aunt Gertrude in her new hat.

Early Brownie camera (courtesy Wikipedia)

Like Hendrix, Pollack was inspired to reduce painting to its essentials. No longer tied to traditional forms, he experimented with materials and abstract ways of expressing his conflicting emotions through paint and canvas. The image hanging in a gallery is only the output of one performance, echoes of the motions, grunts, and groans as he literally attacked and embraced his media.

Today’s young computer engineers and designers pecked at the keys of a computer long before they could read the letters on the keys themselves. There’s no sense of intrinsic amazement when they sit down in front of a PC. The banality of computing, coupled with cheap, disposable microprocessors like Arduino, is leading to a new generation of devices that don’t look like traditional computers.

IBM 5150 (courtesy Wikipedia)

Products like Supermechanical’s Twine are the Woodstock guitar solo of the information age.

Twine is a wireless module tightly integrated with a cloud-based service. The module has WiFi, on-board temperature and vibration sensors, and an expansion connector for other sensors. […] The Spool web app makes it simple to set up and monitor your Twines from a browser anywhere. You set rules to trigger messages — no programming needed. The rules are put together with a palette of available conditions and actions, and read like English: WHEN moisture sensor gets wet THEN tweet “The basement is flooding!”

 

- Twine sales pitch

 

Twine

Stripped of the need for affordances such as keyboard or display, Twine expresses a single emotion: apprehension. A Twine on your basement wall embodies your fear of flooding. The Twine on your washing machine is your guilt for letting laundry sit between cycles. It may not be as beautiful to the ears as a Hendrix guitar solo or as visually stunning as a Pollack, but the essence is the same.

This is an amazing feat. It also shows the challenges facing chip designers who want to push the limits of how fast chips can go.

For some reason — perhaps I’m feeling overly nostalgic this week — this story reminded me of the Commodore 64. For those who don’t remember this darling of the 8-bit personal computer era, the C64 ran at a paltry 1MHz, or 1/8429 of the speed those AMD engineers achieved. But the clock speed wasn’t just a limitation of the 6510 processor inside.

Indeed, the processor actually ran at “approximately 1MHz”. In many early machines, the crystal oscillators for generating the video signal also were responsible for the system bus. Steve Wozniak started the practice of using a 14.31818 MHz oscillator in American Apple II’s, which is 4x the 3.579545 MHz frequency needed to make an NTSC color TV signal. Divide it by 14, and you get the CPU clock speed of early Apple II’s and the Commodore 64: 1.023 MHz.

In both the Apple II and the Commodore 64, the video chip and the CPU share the same memory, because RAM wasn’t exactly cheap back then. The RAM essentially runs at 2MHz, with the CPU and video taking turns accessing it. Now, the 4116 DRAM chips in the Commodore 64 are designed for a 375ns cycle, or roughly 2.6MHz. So you’d need to replace the RAM with chips designed for faster clock speeds.

But then there’s also the matter of keeping everything in sync. In its built-in C64 compatibility mode, the Commodore 128 can actually run its CPU at 2MHz. But enabling this feature disables the display. A faster CPU can’t share the RAM with the video chip. And since the CPU clock is derived from the color video signal, the fastest you can clock the CPU is 14.31818 MHz unless you turn things around and drive the colorburst clock from the CPU clock, instead of the other way around.

What a mess! Maybe the big, exciting thing about the 8 GHz processor experiment is the proof that we’ve decoupled all the clocks inside our computers, that we’ve made crystal oscillators so cheap we can throw a bunch inside one computer. We can use dangerous liquids to clock our processors to insane speeds yet still have video and access RAM!