ROM Computer Magazine Archive ROM MAGAZINE ISSUE 7 — AUGUST/SEPTEMBER 1984 / PAGE 13

Bill Wilkinson

Interviewed By Peter Ellison

    Q. When did you first become interested in computer programming?
    A. The first time I ever wrote a program was when I went to Carneghy Tech back in Pittsburgh, Pennsylvania. I never even took a programming course, but my roommate did. While I was goofing off, not doing my engineering work, I auditted his course and wrote a program on a computer then. I didn't really get interested in computers until '69 or '70.
    Q. When did you get your first Atari?
    A. We actually didn't own an Atari until after we started the business, "OSS". The entire time we did all the work for Atari, we never owned one. They delivered us one and the disk drive didn't work so we gave it back to them and they never gave us back another one.
    Q. What was the first Program which you worked on that sold commercially and to what company?
    A. When I worked for Aero Jet Electric System I did some work on the Satellite Systems.
    Q. Who worked with you on Atari Dos?
    A. There have been a lot of things attributed to me that should not be have been. I've tried to say that in my books, but people don't necessarily believe that. I didn't have much to do with Atari Dos. A fellow by the name of Paul Laughton wrote virtually all of Atari Dos, incidentally he also wrote Apple Dos. This was done through a company named Shepardson Microsystems.
    As I look through the Atari Basic code, I guess I ended up writing more of it then. I thought I did at least influence it more than I thought I did because I see a lot of my comments in the floating point code that I had forgotten I had written. That was about all I was really involved in. The one thing I did do was I was involved in the negotiations for the contract, and I was then one who did the spec for it. So if you want to blame someone for the spec on it, you got to blame me. There were several things done in the final version which were not in the spec and there were some things left out that were in the original spec. Not the least, for example: I never suspected that we wouldn't be able to input into a string array. I didn't think you wouldn't be able to say INPUT "A^3". That was a fallout of the fact there were two people working on the BASIC, and input wasn't done by the same person who did lit. So things just didn't get done. There were quite a few little things that slipped through the cracks of Atari basic.
    I don't like the way the length of the string is defined in Atari BASIC. In Cromemco BASIC the length of the string is defined by the number of characters, up to and including the last null characters. So if you attach nulls onto the end of a string it will automatically shorten. Obviously that is a better definition than the Atari version. Inputs should have been able to handle quotes, an oversight.
    After we did the Atari stuff, we managed to keep ownership of it and Atari got the nonexclusive right to it. We decided to try marketing the stuff on the Apple also. Shepardson never wanted to put any money behind it, and that's where the name Optimized System Software came from. We felt we needed a name with a little more pizzaz then Shepardson Microsystems. That was a resounding failure, because within a year we sold about five hundred copies on the Apple. That's when we were taking on projects in which we got in a little bit over our heads. Basically, Shepardson was tired of having people work for him and we weren't able to agree with one another anymore and that was when we all said, "Bye." Everybody then broke up. I thought there was still some value in the stuff we were selling at OSS software and I negotiated with him for the rights to it. Basically that is where OSS came from. We kept the same name, it was easier than trying to invent a new name. This was in 1981.
    Q. How big is the programming staff and do you have outside programmers?
    A. That all depends on what day of the week it is. We're a pretty small company in general. We've got 15 people, and that includes shipping and everything. How big the programming staff is, depends on how many are answering technical questions on the phone that day. If you're answering technical questions, then you're not a programmer that day, are you? But actively working on programming, there are between 3 and 6, depending on the day of the week.
    Q. Do you get a chance to oversee most of the programs?
    A. The trouble is that I get myself involved in too much around here. So the answer is yes. Basically, for about a year and half there, we didn't produce any products ourselves, we bought them from outside sources and things like that. We're just now starting to get back into producing some new products. It took a while to build up to that point. When we were really small, the two or three of us who were involved with it could write programs and answer phone calls and everything else all at the same time. Then when we got a little bit bigger it took all of our time just to answer the phone. Now you get up to the point where you can hire someone to just answer the phone while you go off and program. Its kind of a pleasant feeling, I hope that it lasts.
    To answer your question about having outside programmers, the answer is "Yes." Not people that normally work here, but do so from time to time when they come in to test final versions. Action! was done outside of us and so was MAC/65.
    Q. When were you asked to write for COMPUTE?
    A. What happened was I saw an answer in "You ask the readers" or "Feedback" or something like that and it was an absolute flat out wrong answer. I called them up and talked with them for a while and said that they should get someone who knows something about the Atari machine to answer the questions, rather than just guessing. They said "Great, you've got the job."
    Q. How long did it take to create BASIC XL and how many programmers worked on it?
    A. Steve Lawrow and myself were the only two who worked on it. We started from Atari BASIC, and I mean that literally. The reason we did it was because when Steve was in the process of doing MAC/65 he had learned some tricks on how the syntax stuff worked and was anxious to try them in basic. One of the things he did first of all was to produce an integer version of BASIC A+. BASIC A + was a thing we sold on the Apple and then we converted it to the Atari. I shouldn't say simply because I added all the Player/missile stuff. Kathy O'Brien, who was originally one of the other persons who worked on Atari BASIC had added a lot of the stuff before I even got hold of it.
    Steve did an integer version that was faster, but was not as fast as we had expected. It was at that time we started talking it over. What can we do to speed it up? It didn't turn out exactly like we had expected, because Steve doesn't take directions very well, but turned out pretty close to what I was hoping for. Basic XL took about six months, which is interesting because Atari BASIC only took two months in the first place. That was because there was a little more forethought put into it. Basically, Atari Basic was written with the attitude of `don't worry about it, we'll put it in later'. We never intended Atari Basic as specified and as delivered in the original machines to be the final version. We always intended it to be a 16K version. And Atari only through reasons of marketing, decided they didn't want a 16K version and instead they went off on a microsoft Basic. The result was a microsoft Basic that was kind of inadequate, a year and a half to two years later than we would have their 16K BASIC. There is a little bit of sour grapes on my part there. I suppose in one sense it's good, because OSS wouldn't be in existence if we'd done the 16K BASIC, because then we wouldn't have had anything to sell. On the other hand, I feel that Atari really blew it because at one point in time they were offered what was equivalent, BASIC A+ for $35,000 flat. Instead they spent at least a million in promoting Microsoft BASIC.
    Q. How did Clinton Parker come up with the idea of Action!
    A. The reason he wrote it, he told me, was because he bought this Atari (Couldn't afford anything better than an Atari 400), and liked it, but couldn't find a decent screen editor. He wanted a full screen editor and he had written a full screen editor on some of the large mainframe computers. He said, `The hardware is here, why isn't the software here?' So he wrote a screen editor. He found out that it only took up three to four bytes of code and said, "Well, I got this 8K cartridge, which he also built from scratch, what the heck am I going to do with the rest of the cartridge? I know, I'll build a compiler."
    Q. Why did he choose OSS?
    A. First he came to us, but at the time we were looking at the ABC compiler and Clinton wouldn't tell us enough about the Action! Compiler for us to understand how radically different it was. He was being secretive. We had already signed a nondisclosure agreement and had pretty much decided we would pick her up (market it). So we told Clinton, "We can't tell you what we're marketing but it sounds similar to what you're doing." And he then said, "No it isn't." So I said, "Tell me more about it."
    Then the fellow who wrote the ABC compiler changed his mind and decided to market it himself and in the meantime Clinton had taken Action! over to Eductional Software. At that time Educational Software was having some financial trouble. Clinton realized the one thing that would be needed in Action was a gigantic manual, and a promotion, and he didn't even understand the thing or what the language was even for. So Robin at Educational Software sent Clinton back to us. Then when Clinton showed it to us we went bananas because it was the first time we had gotten a chance to see it. We saw it as an 8K cartridge, and believe me, even as an 8K cartridge, it was impressive. At that time you had to have all the libraries on the disk and the editor didn't have any of the power it does now.
    Q. How long after that did you get it into production?
    A. We saw it in January or February of last year, '83, and we managed to get an informal agreement with Clinton by the time the '83 West Coast Computer Faire was on. The first day of the Faire we showed it as an 8K cartridge and by the second day we showed it as a 16K cartridge. At the same time that was happening we were also working on BASIC XL, and the whole reason for the Bank Select cartridge was because we wanted to be as compatible with Atari Basic as possible. The only way to do that was to have a Bank Select cartridge. It had to fit in 8K bytes of ROM. Otherwise our market would be zero. We had the first Action cartridge made about mid April but we didn't start shipping until August. We had expected to have it out at the end of June but we missed it, needless to say.
    Q. What new programs at OSS do you have planned for in the near future?
    A. We're coming out with some new products but there is nothing spectacularly new or different, they are just what we'd like to think of as high quality. We have dropped our plans for the $50 word processor. The Atari market doesn't seem to care about another $50 word processor because there are too many of them out there already. Instead we're going to come out with a $129 word processor with a lot more features. The question is how can you sell a $129 wordprocessor when you can't sell a $50 one? We'll sell many, many fewer of them, but the people who buy them will know why they are buying them. This word processor has got most of the features of any other major machine word processor. It will also be in a Bank Select cartridge.
    Q. What do you think of the Atari XL series?
    A. There's so much hidden in the machine that has not been made public yet and I don't know what to think. The 800 XL has the finest operating system in ROM than any other machine on the market today. It is smart enough that when you turn it on it goes out to check to see if there are any parallel devices plugged in. If there are they are automatically hooked into the device driver chain. Everytime you call SIO for any requests, it checks to see if any of the parallel devices want to service the requested step.
    Q. What type of Atari do you use when you're programming?
    A. I'm not really fussy, but most of the time I've been using the 800 XL, not because I think it makes any difference in the type of stuff I do, but only because it's easier to carry around. One rule we have around here is: "If it doesn't work on all the machines then you did it wrong."
    Q. Did any of the programs from OSS need to be converted for the new XL's?
    A. Yes, one. "Bug/65" of all things. Probably the most minor program. Incidentally it will become a standard part of OSS DOS XL from now on as of August 15. Basically we're raising the price of DOS XL to $39.95, but we are throwing a lot more goodies into it. A `Free' debugger into it and things like that.
    Q. What do you do when you're not programming or working for OSS?
    A. The problem with starting up a small business is that you simply have no time for anything else. That has certainly been my situation in the past couple of years. My hobby is music and I have a special wish to some day do a really nice coupling of a computer to a music system. I have yet to see a music system that I want to see on the Atari. It probably can't be done without a light pen. The Atari probably isn't capable of creating sounds the way I want it to. I want to be fair about it. This is not a criticism of the Atari, its just a limitation, but still I believe there is a lot that can be done on the Atari that hasn't been done yet.