by Frank Rose

Frank Rose

Frank Rose and the keyboard of his "power everything" Epson QX-10.

Frank Rose is a contributing editor at Esquire. He is the author of Into the Heart of the Mind: An American Quest for Artificial Intelligence.

It all started when I was on the phone with the president of Rising Star Industries, the California software firm that was programming Epson's soon to be released QX-10 microcomputer. What, I asked, was a rising star? It was Japan and America, he said, the Rising Sun and the Stars & Stripes. The Epson computer and its Valdocs word processing software would be a marriage of Japanese engineering and American ingenuity. And as a professional writer, he mentioned, I'd make a perfect beta test site for the new machine.
    What was a beta test site? I called my friend in Silicon Valley. A beta test site, he said, is a novice who tests a new computer. Alpha tests involve company personnel; beta tests involve random specimens from the general public. I'd be a guinea pig: if there were any bugs in the system-and there would be-I'd be the one they'd bite.
    I called back and said yes.
    Two months later a series of cardboard boxes arrived at the door. Inside was the Epson hardware: an 8-bit machine with 256K of random access memory, using the Z80 microprocessor and running on the TPM operating system (an advanced version of CP/M). At the time I didn't know what all this meant, but I would soon find out.
    When set up, the Epson QX-10 looked like a sleeker version of the IBM PC: lean, low-profile, ready for business. Unfortunately, I couldn't do anything until the software came.
    Another two weeks and the software arrived, but there was no manual, no instruction book, no documentation, no anything. This, I thought, must be part of the test. One of the main selling points of the Epson, I'd been told, was "user-friendliness." This was intended to be a machine you could plug in and use right away, like a stereo or a toaster-oven. I was dubious. But I did know that the Valdocs disk was supposed to go in the A drive and the data disk in the B drive, so I turned on the machine, slipped the disks in and pushed the button marked HELP. The screen filled up with a few basic instructions, and before long I was ready to go to work.
    When the documentation finally showed up, I learned some new tricks: how to delete text by the word or the line, how to get into the "electronic address book" that threatened to replace my Rolodex, how to use the CALL button to do calculations on figures in my text. Over the next few months I learned also to appreciate the clean design of the on-screen text editor and the effortless nature of the keyboard. It was the keyboard that particularly impressed me-a long slab of plastic that encouraged easy access and fluid motion.
    But there was a downside to the beta test experience as well. I was working with experimental software, and one of my duties (aside from suggesting ways it might be improved) was to fall into any hidden traps that might be lurking in the code. These traps took any number of forms and appeared at random. One was simple data loss: I might store a file and find it half a page shorter when I came back to it. Then there was data scramble: sentences and fragments of sentences were liable to turn up in any order they pleased.
    My favorite bug, however, was the one I came to know as the "screen of Damocles": at any moment the text I was working on might be replaced by a screen full of nonsense characters. Once that happened, there was nothing to do but turn off the machine, turn it back on and return to the file as it had last been stored on disk. To call this daunting would be an understatement; but I like to think that for a while, at least, it lent a certain existential quality to my writing.
    Eventually the bugs were cleared up, extra features were added and Valdocs began to work the way it was supposed to-cleanly, logically, efficiently. Only then was it possible to judge the system as a purchaser would, on the basis of what it actually did instead of what it was supposed to do or what it might do if you were unlucky. When that happened, I became aware of a curious fact.
    Like a two-ton Cadillac with power everything, the Valdocs software is so loaded down with luxury items that its performance can be a bit sluggish. Everything works-you can get on-screen italics, and long file names, and two print modes, and on-screen graphics, and on-screen underlining, and a multitude of block commands, and electronic mail at the flick of a button-but it isn't going to set any speed records. Was this the Rising Sun and the Stars & Stripes, I wondered, or was this Detroit?
    That said, how do I feel about my road hog? I love it. It looks great. It makes me feel good just to sit behind the keyboard. Its 256K of main memory and 376K of disk storage capacity makes it big enough to push other machines off the road. (I've even started talking about memory capacity the way hot rodders talk about horsepower.) I know its text editor like the back of my hand. And just last week I got a commemorative plaque from Rising Star informing me that, as a beta test site , a part of me " resides within its code." While searching for a spot to hang the thing, I wondered: Does this mean I can load it in and go visit myself?


A computer is a piece of equipment that can be made to follow a predetermined plan or to obey a spell you cast on it. That's the good news.
    The bad news is that creating such a plan, casting such a spell, is much harder than you'd think. The mechanics of spellcasting-or, as it is called, "programming"-becomes the center of people's lives. All kinds of clever fellows keep inventing new spellcasting systems, so the hope is dim for standards, stabilization or general transportability between all the different computers.
    Why such fascination with the mechanics of planning and casting spells? Because so many new possibilities keep coming into view. Only gradually are we learning that the art of programming has as many styles and systems of composition as, say, music.
    Each person's viewpoint is different. Practically every spellcaster wants to do things a different way, and a lot of them eventually get around to designing their own spellcasting systems-or, as they are called, "computer languages." The new language builders want to accommodate a new range of variations, thinking they see a way to reduce the complexity of design and arrangement. So they invent a method, a language, that opens up, they think, rich new possibilities. And, indeed, frequently it does-and now the computers can be made to do new things, or combinations of old things with a less difficult spell, and its users develop a style of their own.
    If the new language is successful, enthusiasts and usages proliferate. The new possibilities are tried in combination after combination, and assembled by enthusiasts into new skyscrapers of combinations. And inevitably there are difficulties and complications as these detailed edifices of plan-structures grow. And the complications increasingly frustrate some of the enthusiasts, and there we go again.


Return to Table of Contents | Previous Article | Next Article