I invented the iPhone in 1977

Pocket computer

Did I draw this? I think I must have. I can’t imagine how.

No, really. I did.

You can read about it here: David Chapman, “The Pocket Computer,” Interface Age, December 1977, pp. 72–74. (The full text is below on this page, too.)

I had forgotten about this for decades. I found a copy in the jumbled mess of my mother’s house after she moved into assisted living. It took me months to bring myself to read it; I was expecting it to be hideously cringey. In retrospect, it is naive, but I was fifteen, so maybe my former self should get cut some slack. Also…

It is prophetic. I mean, I nailed it, pretty much. I radically underestimated how long it would take for the hardware to be feasible, but the UX concept was just about right.

And, as far as I can tell from a couple hours of internet research, this was the first time someone had suggested anything similar.

The Wikipedia “Pocket-sized computer” article notes a first use of the term “pocket computer” in Blondie’s 1978 song “Picture this.” (Did Debbie Harry read Interface Age?) More seriously, the first actual pocket computer was developed by Sharp and sold in 1980 by Tandy; it was basically a programmable calculator, and not much like my 1977 fantasy.

Tandy Pocket Computer

Image courtesy Jcassara86. I wonder what that P=NP button did?

The first computer that was mostly screen in front was the 1993 Apple Newton. You could write directly on the screen with a stylus, either “zapping characters on a pseudo-keyboard” (as I suggested) or using software handwriting recognition (as I also suggested). It was more a tablet than pocket sized, it was too expensive, and the handwriting recognition didn’t work well, so Steve Jobs killed it. It did demonstrate near-feasibility of the concept.

Apple Newton and iPhone

Apple Newton and iPhone, courtesy Blake Patterson

The 1996 PalmPilot was small, affordable, worked, and was commercially successful. Over the next decade, many companies steadily improved the product category.

The first pocket computer I owned, in 2000, was an iPAQ 3670. That was the first device to get wide-area wireless internet access, as far as I can remember. It ran Microsoft’s surprisingly good Pocket PC operating system.

iPAQ running Mathematica

Image courtesy David Kirkby

When I got it, I remember vaguely thinking “didn’t I invent this when I was in high school?” But I didn’t have a copy of the article.

It was, in fact, very similar to what I had dreamed up in 1977. I was only off by about twenty years.

It was, in fact, more similar to my dream than the iPhone was. (So I lied. I didn’t invent the iPhone; I invented the iPAQ.)

The iPhone’s main innovation was a better touch screen. That made it possible to eliminate the stylus used in previous pocket computers, which somehow seemed geeky to non-geeks.

Selling the iPhone as a phone, not a “personal digital assistant” or “pocket computer,” made it hugely successful. Even though, even then, even for non-geeks, “phone” was a minor part of its use; and even though phones running the Pocket PC operating system had been around for several years.

I often miss the stylus. Finger-typing on an iPhone sucks.

The article

Thank goodness for The Internet Archive! They have a text version of the article as well as the image scan. I’ve copied it here.

In the special issue introduction, the Interface Age editor noted (page 5):

THE POCKET COMPUTER is yet not in your pocket. It is still over that ill-defined border between science-fact and fiction. However, all the elements are extant and production is imminent. When? Neither David Chapman, the author, nor the staff of INTERFACE AGE got a very good reading from the crystal ball. When we urged it on to uncloud and read out an answer, it momentarily flashed “Soon.” How soon is soon?

The Pocket Computer

by David Chapman

Computers get smaller all the time; but it seems that there is a downward quantum jump in physical size that occurs near the beginning of each decade. First there were the mechanical analog monstrosities of the For- ties, then the tube and relay kluges of the Fifties, the transistor and SSI systems of the Sixties, and now the LSI microcomputers in the Seventies. What’s next?

Certainly further miniaturization would be desirable. The microcomputer, though relatively small, is far from portable by the time you add memory and peripherals to it. It can not be carried around the way a pocket calcula- tor can, which severely limits the times and places it can be used. The difference between a hypothetical pocket computer and the micro is like the difference between a modern calculator and an adding machine: one is of universal utility; the other is restricted in use to a small group, “a computer in every home” notwithstanding. To bring the computer to the people, I suggest a computer in every pocket!

Let us now look at how a pocket computer might be made commercially practical. The main restriction on the miniaturization of the microcomputer is no longer the CPU itself, nor even the memory, but rather the I/O devices, particularly the keyboard and the CRT, the only tube in the IC age. This limitation can be overcome with technology available now — although the development time will probably delay the introduction of the pocket computer until the early 1980s.

The pocket computer, as I visualize it, would be similar in size and shape to the pocket calculator, a time- and consumer-tested design, about six inches long, four inches wide, and three eights of an inch deep, with a plastic case and simple instructions on the back.

The general design philosophy should be to make it instantly usable to anyone who can read and write. Instead of a keyboard and seven segment readouts, the entire front would be covered with a flat LED, liquid crystal, or plasma screen, like the ones on PLATO terminals. Prototypes of all these designs have been in existence for several years since they were developed by the television manufacturers as possible substitutes for the CRT. This screen would probably be the most expensive single item in the computer, not including software. At first, low definition black-and-colored versions with perhaps 256x256 dots would appear; then as prices drop and the technology improves, full color, multi-intensity screens might become available, with definition as high as 1k2 .

The screen would undoubtedly consume a large amount of current, which would necessitate large storage batteries; but as the inside of the box would probably be otherwise largely empty, this should cause no problem.

The processor, memory, and interfaces would be on a chip or chips glued to the back of the screen. A possibility is to make the chips plug into each other in some way so that one could add a new 32K memory chip or a line printer interface to one’s existing system. A minimum configuration, which would come on a single chip, might include, say, a 16-bit microprocessor with BASIC, APL, and TRAC on ROM, perhaps with an assembler/text editor/linking loader and a USR or similar machine language subroutine call to allow special routines that are not possible in the high-level languages; 32Kx16 RAM for program and data storage; screen, lightpen; and interfaces to disc drives, lineprinter, and what-have-you for use when the computer is being used at a desk.

An important component of the pocket computer would be the lightpen, the main input device, which would replace the keyboard. It would be attached by a wire, Koil Kord™, or radio link to the computer, and would be similar in operation and construction to present-day lightpens. In primitive versions of the computer, it would be used to “zap” letters displayed on the screen in a pseudo-keyboard; but as pattern recognition techniques get to be more sophisticated, direct handwritten input should become possible (see Figures 1 and 2). Handwritten input would be extremely fast, and very easy to use.

Another input possibility is voice. This is still in the development stage, and is not yet sufficiently sophisticated for our purposes. The essential problem is that the same string of phonemes, or sound units, may mean different things in different contexts. For example, the words “I scream” and “ice cream” are virtually identical in sound, but are easy to tell apart when someone asks you “do you want some ice cream?” or says, “when I hit my finger with a hammer, I scream.” This implies that the word-recognition software has to interact with syntactic and semantic parsers, which is rather complicated.

Voice output, on the other hand, though technologically a “solved” problem, is very limited in its application because it requires the user to remember what’s been said.

The main output device would, of course, be the screen. Full graphics as well as characters are possible since the hardware would presumably bitmap the computer’s RAM onto the screen: that is, each bit in the bit-mapped memory would correspond to a dot on the screen; if the bit is “high” the dot lights up; if “low,” it is dark. However, the addition of graphics generation to the other input/output chores of the processor would probably slow it down to an unacceptable pace. Therefore, a slave processor could be used to handle all I/O operations, while the main processor did the “heavy thinking.” The slave would have full graphics routines and I/O handlers on ROM, with the capability of drawing vectors, arcs, geometrical figures, or whatever.

One problem with this scheme is mass storage. Traditional mass storage is very bulky; floppies are an improvement, but there is no way they are going to be made pocket-sized. I see two possibilities here. The first is to use some type of miniaturized card or tape reader/ writer like those on the SR-52 and HP-65. These could be either magnetic, or, possibly, laser-based optical devices like video discs. The other possibility is to use magnetic bubble memory, which is limited in the amount of information it can store.

By the early 1980s when this becomes commercially possible, natural language processing may have advanced to the point where computers could be programmed in a subset of English or a synthetic language like Loglan. This may or may not be desirable. Although natural languages are hardly optimal for computer programming, they might allow those who don’t want to learn “computerese” to program simple applications for which they can not find a “canned” program.

The pocket computer might be a good machine to implement some of the far out “reactive grand national computer network” type of ideas that are in the air, like Xerox PARC’s Dynabook, Ted Nelson’s Xanadu system, and SRI’s NLS. This is simply because the pocket computer is, I believe, the trend for the future.

The technology for all of the hardware ideas I have mentioned is available NOW, but someone (Zilog, Intel, et al., are you listening?) has got to make it affordable. To be practical, the maximum price that would make a pocket system competitive with the micros is about $1000, and a $25.00 model would certainly be much better.

Navigation

You are reading a metablog post, dated August 13, 2020.

The next metablog post is Hypertext books vary in shape.

The previous metablog post was A slice of The Eggplant.

General explanation: Meaningness is a hypertext book. Start with an appetizer, or the table of contents. Its “metablog” includes additional essays that are not part of the book.

To hear about new content, Subscribe by email subscribe to my email newsletter, Follow Meaningness on Twitter follow me on Twitter, use the Syndicate content RSS feed, or see the list of recent pages.

Click on terms with dotted underlining to read a definition.

The book is a work in progress; pages marked ⚒︎ are under construction.