IMSAI - 'I Must Save Apple Immediately!' A very unique historical IMSAI

     The "Blue Box" pictured here is obviously not one of the more famous made by Woz and Jobs for prank calls to such places as the Vatican.  But it did play an important role in the early days at Apple, and maybe even affected the fate of these afore mentioned pranksters.

     On November 14th 1977 Cliff helped me move this IMSAI from the car, along with a home made video terminal and other peripherals, to my desk very near the front door of what was the Engineering lab of Apple Computer Incorporated.  Apple was then located behind the Good Earth restaurant on Stevens Creek Blvd. in Cupertino California.  It was my first day at Apple; Cliff would be starting two weeks later.

     Cliff and I had been warned that Apple was in a round of financing to continue operations and expand production.  If the financing didn’t come through, our employment might be very short.  Among the immediate promises to the investors was that Apple would ship a “floating point” BASIC in January, and a floppy disk drive by June.  Applesoft was to be that floating point BASIC.

     We were met with the skeptical eyes, and a not a few derisive comments, of one Randy Wigginton (Apple Employee #6), a 17 year-old kid that worked "part time" at Apple as a Software Engineer.  I say "part time" because he was supposed to be (from a legal and his mother’s standpoint)… but I'm quite sure he was there much more than 40 hours a week.  Randy at the time was working on getting Applesoft up and running, starting with a buggy mess of a licensed version of Microsoft BASIC that had been converted to run on the 6502 processor.  Randy had to deliver a working version, with Apple enhancements, in January.  Randy was the sole programmer on the project, so you can imagine the pressure he felt.

     Randy did his work on an ancient Teletype machine with roll paper and a paper tape punch/reader.  Connected to the Teletype was a 110 BAUD (yeah, 110 bits, each and every second) modem communicating with a timeshare service running a cross-assembler written in BASIC. A full assembly of Applesoft took about two hours, if no one else with higher priority was using the system.  Everybody else had higher priority.

     Randy had piles of the paper roll segments he had printed out with various hand written notes and changes he'd made to the code along the way.  He also had a couple of full printouts of Applesoft, in various stages of modification, from a line printer at the timeshare company.  Randy would pick these up every few weeks, rather than trying to print the whole of Applesoft on the Teletype.  At 110 BAUD, it would have taken several days to do that, assuming that the timeshare service was actually available for that duration!

     I had wanted our home-built, custom programmed IMSAI at Apple so that I could work with something of which I was familiar- kind of like a kid needing his blanket - rather than having to immediately learn how things were done at Apple before I could be productive.  I had written a crude disk operating system, a text editor, and cobbled together a cross-assembler (from an available 8080 assembler offered by Processor Technology Corp.) in the weeks before starting at Apple in order to learn the 6502 assembly language.  My first week at Apple was spent mostly trying to figure out how to get the cross-assembler generated code out of the IMSAI and into an Apple, where it could be useful.  With Cliff's after-work assistance, an Apple prototype "parallel card" and driver was quickly put together to transfer text addresses and data from the IMSAI into the Apple's monitor program, and thus into the Apple's memory.  From there, the code could be written into the little 256-byte ROMs used for Apple's peripheral cards using a card and program that I believe Wendell (employee #16) had designed.

     I had been hired as a programmer, though I had no college training in the art.  Like Woz, Randy, and Chris Espinosa, I was a hobbyist, or perhaps more accurately: a hacker.  I think the only reason I was hired was because Apple was set on hiring my brother and he offered the two of us up as a package.  Besides, I was cheap - if I didn't work out, they could ease me out after Cliff had started, without wasting too much money!

     I could understand Randy's distain for the big blue box beast.  After all the money (and work) Cliff had poured into the IMSAI, it still didn't have color graphics, like the Apple; it wasn't friendly with a nice case or built-in keyboard, like the Apple; nor was its processing speed faster than the Apple, even though its clock speed was double Apple's.  It had cost three times that of the elegant little Apple, countless hours were spent to build it, and it weighed a ton.  To even get the IMSAI going, a sequence of binary codes had to be entered using the 25 toggle switches on its front panel.  The Apple was ready to go with a single flip of the power switch.

     The IMSAI did have three things that the Apple didn't have:  a floppy disk drive, a paper tape reader, and a 6502 cross-assembler program.  These features would prove to be crucial in early December when there was a crash on the timeshared mini-computer that Randy relied upon to develop Applesoft. 

     The timeshare system failed and the magnetic tape that held Randy’s work became unreadable.  Randy had lost his efforts for the few weeks since it had last been backed up.  The timeshare service dutifully mounted the previous month's backup tape, which also promptly became unreadable.  So, the service went back another month, to when the first work on Applesoft would be accessible.  NO GOOD!  It turned out that what had failed was the Tape Drive.  The failure was that the Write/Erase mechanism was permanently engaged; meaning that as they put on each backup tape it was erased.  Randy was back to square zero.

     When Cliff and I arrived at Apple that December day Randy was growling, pacing, yelling on the phone during calls to the timeshare service, and generally having a bad day.  This was not the ranting of a 17 year-old boy, but the full throated anger, frustration, and verbal threats that could be expected of anybody whose hard work was being systematically destroyed by people he had no real power over.  Four months of work gone, everything due just a few weeks away.  And then things quieted down.  Randy was seated in his chair, looking at the paper tape that was the Microsoft BASIC with which he'd started.  And then he hoarsely asked me, "Do you think your assembler could handle this?"  I think this was the very moment when I was potentially, provisionally, not just a BOZO in Randy's judgment.  I might be a useful BOZO. I replied: "I'll have to make a few modifications, but yeah, I think it could."  I didn't want to be judged a mere BOZO any longer.  Useful was a step up.

     Changes were necessary to get the cross-assembler to handle such a big job.  After all, the source assembly language code was many times the size of the IMSAI's memory and just the binary code output was nearly 10 kilobytes!  My assembler program worked only from source that was preloaded into memory, so the first task was to break-up the Microsoft source into digestible chunks, in the form of disk files less than 18K each, as we manually pulled the paper tape through the IMSAI's reader.  The source ended up taking multiple 5 1/4" 80K diskettes to hold it all.

     In order to then assemble Applesoft, I needed to add a directive to automatically read the sequence of files, as well as pause dutifully to wait for the next diskette to be inserted when the file couldn't be found on the current one.  You see, the IMSAI, at the time, had only one disk drive.  And then there were modifications to allow for Labels and Symbols that were not tolerated in the Assembler, which happened to be ubiquitous in the Microsoft source.  Lastly, before our first successful assembly, was dealing with the shear size of the resultant Symbol table generated – we needed more memory.  Adding a S100 16K memory card to the IMSAI solved the problem. Fortunately a bare memory card was available immediately at a computer shop in the valley.  Cliff fetched it and put it together using RAM chips he already had, within a couple of hours.

     All of this trauma, and much of the recovery, happened without consultation or even awareness by Apple’s upper management.  We didn’t need no stinkin’ badges.

     By the next week Randy was banging away on the hand-built keyboard video terminal that was attached to the IMSAI.  He had his notes and printouts, so it hadn't been a total loss.  And there was a huge upside to switching to the IMSAI that wasn't readily apparent to Randy when it, and I, arrived a month earlier.  The IMSAI could assemble Applesoft in just 6 minutes or so, even though it required someone to swap diskettes as it ran. Not just that, but within another couple of weeks Mike Scott (Apple's President) took in the situation and said "We need a printer."  By January we had a brand new, louder than all Hell, Printronix 300-line per minute printer at our beck and call!  Applesoft could be assembled and printed in entirety in less than an hour.  The Applesoft tape version and Applesoft card shipped according to original planning; Randy had delivered on time!  The investors were happy, the boss was happy, and Randy was happy.

     There was a downside, too.  I no longer had exclusive access to my blanket.  Oh, well.  I only needed it for piddley little 256 byte peripheral ROMs for the Communications Card (Wendell's code), Printer Cards, floppy disk drive boot ROM, High-speed Serial Card, and a few larger items, like the first Apple DOS and DOS 3.2 (the first stable version!).  Over time, as the staff grew, a couple of North Star systems were purchased to also run the cross assembler. Some of us had started working on the next Apple and the Pascal team had their needs, too.

     [A little over a year later, most of us used "Randy's Weekend Assembler"  (despite all that had happened, Randy still didn’t “like” the IMSAI), that was later expanded and improved by John Arkley before being released to the world as "EdAsm".  Imagine, at last, that we could actually use an Apple II to develop code for the Apple II (and Apple III, Apple IIc, and Apple IIgs), despite its original shortcomings.]

     But there were other upsides, as well, that should be mentioned.  Cliff was offered, accepted, and paid a tidy $500 per month rent for the several years that this IMSAI was in use at Apple (some large assemblies were still done on the IMSAI).  Not a bad return on an original $5000 investment.  And I was definitely not a BOZO and was allowed to keep my job at a tiny computer company in Cupertino, California, as a Software Engineer.

     Now you know how a rival’s computer system was employed to save Apple by keeping investors happy, and affect the destiny of a couple of blue box pranksters, in those very early days.  And the IMSAI offered here is the very one that did so...