Dan Bricklin and his friend Bob Frankston founded Software Arts in 1979 to produce VisiCalc, the first electronic spreadsheet. Spreadsheets used to be made on paper. As a student at Harvard Business School, Bricklin thought how convenient it would be if they could be made on desktop computers instead. He wrote a prototype in Basic over a weekend, and then he and Frankston set about turning it into a product. When their first release shipped in October 1979, it ignited the personal computer software revolution. VisiCalc was the "killer app" for personal computers: businesses bought Apple IIs just to use it. Unfortunately, VisiCalc was not produced by a company organized like a modern startup. VisiCalc was developed by Software Arts, but distributed by Daniel Fylstra's Personal Software (later renamed VisiCorp), which paid royalties to Software Arts. Friction between the two culminated in a lawsuit in September 1983--just as Lotus 1-2-3 hit the market. The distraction proved fatal. As a business, Software Arts's fall was as fast as its rise, but it had more influence than many longer-lived companies. Bricklin and Frankston's ideas live on in all the software we use today Livingston: How did you know Bob? Bricklin: I met Bob when I was a freshman at MIT. I was working in the labs as my student job--because a really good way to learn an area in college is to work on a real project in one of the labs. I worked at the Multics project, which was a major project in the history of operating systems. Out of it came the Unix system and the 386-style chipset and a whole lot of things about how we do software and operating systems today. The first job I was given was to make some modifications and finish the work of this other guy, who had just graduated, in his bachelor's thesis. And that was Bob Frankston. 73 Dan Bricklin Cofounder, Software Arts 5 CHAPTER Photo by Louis Fabian Bachrach Bob's thesis was a project called Limited Service System. We used time sharing then; we all shared the same computer over a terminal. The Limited Service System was a way to throttle your usage so that nobody would use more than a certain amount, so they could just give it away for free and know that nobody would hog more than a certain percentage--because this one system was being shared that could handle maybe 50 users or 100 users, and this is for the whole campus. Many of us working at that project were undergraduates or graduate students. Those of us who were young and single would get together socially, too. Bob had a car and lived off campus. He would drive us places, so we all got to know Bob really well. Bob and I always wanted to found a business together. We both had parents who were entrepreneurs, so the idea of running your own business was a normal thing. There are people who come from backgrounds where they're used to working for a company, and they couldn't dream of doing it themselves and not having that safety net. When your parents and family are entrepreneurs, you know it's nothing special. I worked at big businesses and I worked at small businesses beforehand, so the idea of starting your own business was just a normal thing. Bob and I were sort of looking for years for something to go into business with together, and clearly it would be in computers. It's not uncommon to get together with friends that you meet in college. You see that in a lot of startups. The other advantage of the two of us being friends, and not just business associates, was that a lot of the structure of our deals together was based on friendship and not on other things. The friendship was stronger than a lot of the business stuff. So even though we came to odds about things, even though there might be a "Well, did you do more, or did I do more?" because we liked each other and had a relationship, we were able to keep that from messing up the business. We'd be arguing all the time about stuff, but, on the other hand, we have a strong friendship that still continues. Twenty-five years later, we're still close friends. So that was a help, because we didn't have to think, "Do you get 35 percent and I get 65 percent? How are we going to do this?" So many things were just, "We'll just do it 50/50. I'll do this one, you'll do that one." That did make a difference. Also, because we knew each other, there was a lot of trust, which you need, especially in families, because family money was involved when we started the business. Livingston: Is that how you first got money to start a company? Bricklin: We first started on our own. I was in business school, living as a student on loans and savings. Bob was actually working as a consultant, so he was getting money. We went through very little money to begin with, because we used time sharing to do the programming. It was done on a separate computer that you would log into, and then the resulting product was downloaded into an Apple II we borrowed from our publisher, and then it was tested. 74 Founders at Work Bob already had equipment. He had an acoustic coupler modem and a terminal to edit on, from his other consulting work. So we just had to pay for the time-sharing time, and he used it late at night, when it was cheap. I mean really late. Basically, he slept during the day. Livingston: That was at MIT, right? Bricklin: We used MIT's Multics system, the one we worked on. Livingston: Did they mind? Bricklin: No. We paid for it. Luckily it took a few months to be billed. So money went into that, and Bob had some money and was able to pay for it. Eventually we borrowed some money from relatives, because we wanted to buy our own computer. We borrowed money from a bank and from relatives, and we bought a Prime minicomputer, which had an operating system based on the ideas of Multics, done by people who used to work at Multics. We bought one of those of our own, and we sublet space through some other friends who had a business, and that's how we started our business--in a basement. The original business was started in Bob's attic in Arlington, Mass. Livingston: At this point, you had graduated from MIT and were at Harvard Business School? Bricklin: Right. I graduated and worked for a few years, which was important. I had worked for DEC--Digital Equipment Corporation, a big company. Then I worked for FasFax Corporation, a small company. I got to see the differences and see that small companies were just as exciting and just as cuttingedge. You didn't have to be in a big business, which was an eye-opening thing for me. Then I went to Harvard Business School, which was where I came up with the idea. I saw the need for it. But that was coming off of my experience with word processing and typesetting at DEC. I worked in computerized typesetting at DEC because I like practical stuff. My father and grandfather were printers. Out of typesetting, I got into video editing for typesetting, and out of that, I ended up in the word processing group. I was project leader of the first word processing system that DEC did. So that got me into this whole interactive, screen-based, what-you-see-is-what-you-get type system. When I was at business school, taking the experience of what I had done at MIT with interpreters... I worked on the APL system, I worked with Bob on his Basic system; I had done interpreters (in high school I was building interpreters). So the idea of an interpreted language, together with the word processor--and you're sitting there in business school running numbers--the idea of word processing with numbers to me was a natural thing. The traditional way a lot of people think of spreadsheets is as rows and columns, and it really isn't. It's really a two-dimensional layout of words and numbers. If you look at what we had in all our cases at Harvard Business School, at documents you have in business, you have tables of things, but they're organized in a way that is appropriate to the data, and there's a lot of other text, and the text is just as important as the numbers. Dan Bricklin 75 I took this general layout idea of the word processing and computerized typesetting world, together with the calculating world of APL and Basic and stuff, to the needs of business, where you need to be able to ad hoc throw anything together and make changes. That's where the idea for the spreadsheet came from. Then through business school, I met this publisher, Dan Fylstra, of Personal Software, and his partner, Peter Jennings. Dan was a second-year Harvard MBA student when I first met him. When I started programming, he had graduated and was running this business selling software on cassettes out of his apartment in Allston, Mass. He was looking for new stuff, like a checkbook program. I actually prototyped VisiCalc on one of his machines over one of the vacation weekends. I went to his place and wrote a prototype in Basic. Then we started discussing that they would publish it. As MBAs (both he and his partner were MBAs), they understood the value of this thing. They already had a need listed in their list of things they wanted of financial stuff. And they were looking at other financial forecasting tools, but this also would do checkbooks and other stuff. So they knew they could sell it as that; they knew that they would use it. And we made a deal to produce it. I had already prototyped it and said what it would do, but I didn't have time to program it since I was in school. So, since Bob was out of school, he would program it. Livingston: You did it over one weekend? When was that? Bricklin: The fall of '78. Livingston: You just wanted to see if it would work? Bricklin: No, I had been thinking of the idea; I had daydreamed about it. I had actually done a prototype on Harvard's computer system that was available to us as students. As part of the prototyping, I came up with what we have today: the A-B-C 1-2-3 type of thing, the columns and rows ways of indicating things; the idea of having a formula on what we call the contents line that tells you what you're pointing to; moving around where you could move the highlight around from cell to cell--that whole thing. The idea and some of the prototyping had been done. The actual trying it on a personal computer was written in Basic to see what it would feel like. And then we actually programmed it in assembly language starting the winter of '78/'79. Livingston: When you first wrote the prototype that you did in Basic, what surprised you most? Bricklin: I had originally wanted the thing to use a mouse. There was no mouse on the Apple II at the time, so I was using the game paddle and turning it. But the way I was doing it with the game paddles, the cursor was just too unstable. So I switched to the arrow keys, which were much more discrete. I learned some computer things. I had it make a sound every time it recalculated a cell, but it turned out that the making of the sound on the Apple used up three-quarters of the CPU time, because it did it with a timing loop. I learned little things like that. But I saw that it was a useful thing and that it 76 Founders at Work actually felt good and that I could start describing it to some classmates. One of them was also an MIT grad and computer person, John Reese. I would tell him how it was, and he'd say, "Well, Dan, it would be easier if you did this," and I said, "You're right." There was a lot of feedback that way. Livingston: Were you nervous to tell anyone about your idea? Bricklin: No, not those people. Once we started working on it and were in business, yeah, since we thought it was obviously such a great idea. Though we realized it takes forever for it to become big in the world. We didn't think it would be as big as it is now, because nothing had done that in the past, though we thought it was real important. But you always do, as an entrepreneur. Everybody feels that way about what they're doing. You need that drive. And, yeah, we were afraid that Texas Instruments would find out about it and they'd steal the idea. So we were careful; we would have people sign nondisclosure agreements. Livingston: The idea of a startup was pretty new. How did you know what to do first? Bricklin: There were always startups. A huge portion of the economy in Massachusetts came from people who got their start at DEC, which started as an entrepreneurial thing. Then the same thing happened on the West Coast with Hewlett-Packard and places like that. But there was this other business, Personal Software, the publishing company, which was the model that they used of how to do software. This was a different model of author-publisher. We now know that author-publisher is not a very good model. We were the poster child of it not being good. But we set up that way, so when Bob and I made a deal with the Personal Software people in the fall of '78 to produce this product and they would sell it, we needed a business. We incorporated the business on January 2, 1979, and then we negotiated the deal with Personal Software. We were developing the product, but before it was announced we had already agreed on the general terms. The actual specific contract wasn't signed until the night before we announced it at Ben Rosen's conference. We had our lawyer (a general lawyer) negotiating on our side, and we had a publishing lawyer on the other side, I think, negotiating, which wasn't exactly right for software. Our contract ended up having problems long-term. But, it actually ended up being the model contract for many, many software things afterwards, because it did have a lot of interesting stuff in it. Livingston: So VisiCalc was the first to use the author-publisher model? Bricklin: We weren't the first, I'm sure, but in the personal computer software business, we were one of the first. Personal Software later renamed themselves VisiCorp. Dan Fylstra, who was the head of the company, was one of the founding editors or something of Byte magazine. So he was involved in many ways in the publishing business. I assume that his lawyers were from that business too. Dan Bricklin 77 Livingston: You incorporated over your winter break, right? Bricklin: Yes. When I graduated business school, I graduated as chairman of the board with no salary. We announced the product on Monday, and I graduated Wednesday or Thursday, something like that. The first time it was shown to the public was at the National Computer Conference in June of 1979. It had actually been announced privately at Ben Rosen's conference and then shown at the West Coast Computer Faire in May of '79, but only shown to dealers behind closed doors. Livingston: When you were giving these private demos, was there any part of the demo where you just saw the audience say, "Oh my God"? Bricklin: It depends on the audience. We announced it at the National Computer Conference; it was written up by a Morgan Stanley analyst in the summer... you think it would be mentioned in any publication? A business publication? No. Eventually, in a publicity thing about the software publishing that Personal Software was doing, BusinessWeek mentioned it a little bit. And eventually Fortune magazine and Inc. ran stories that we were featured in on the business of publishing software. But, the concept that the spreadsheet as a type of software was available (other than in the personal computing software magazines like Byte or Creative Computing) just wasn't mentioned. I think Forbes finally mentioned it in a comparison of new computers--did it have VisiCalc or not? So it sort of was missed. People who saw it, who needed it, got it. Sorry, no--some of the people who needed it got it. You have to be a person who is able to look at a generalpurpose tool and be able to think, "How would I use that to solve my problem?" Most people are not that way. They look for a tool that is being used already for something close to their problem and then understand what it is. Many people who saw the spreadsheet with an example, if the example wasn't in their field, they couldn't make the leap. Because they're not programmers in their mind. But, if you showed it to somebody where it clicked, either because they understood the general-purpose nature and could apply it to their own needs, or you showed them an example, like financial forecasting or something that they did, and they knew the other tools in the world, they got very excited. If you showed it to a computer person who didn't have those needs, they'd say, "That's kind of cool, but what's so special about that? I could just do it in Basic." Now, there were those that hadn't seen as interactive a computer before, weren't as aware of word processing and some of the other things, and, when they saw it, it really opened up their minds to what you could do interactively with computers. Jean-Louis Gassee, who went to Apple, is one of the people who says that. There were those people--not that many, but enough that it got a lot of people going in computer software. And then there were people--the general public--who thought computers could do everything, and they weren't at all surprised. They'd say, "Well, of course, computers can do so much more than that. What's special?" Luckily for us, the people who funded things--the MBA types got it, the investment banker types got it, because this was something they would need. And that made them get the personal computer. 78 Founders at Work Livingston: Did it drive sales for the Apple II along with VisiCalc? Bricklin: Well, for Apple, yeah. Eventually we could track Apple sales by how many we sold. But the first year we were only selling a thousand units a month. Livingston: Who were the very first users? Bricklin: There was Al Sneider, locally, who was at Laventhol & Horwath, which is an accounting firm, and he started pushing them to use personal computers. They did a lot of accounting for the gaming business. They actually used VisiCalc to figure out how to lay out a casino and where to put which slot machines, I'm told. There were doctors who had bought personal computers because they thought it would be kind of cool, who used it for, I think, anesthesiology calculations in open-heart surgery. We got cards back where people said what they used it with; we asked them in their registration card. They were people who liked technology and were enamored with the personal computer, who knew business. But, as I say, only a thousand units a month. It took a while for people to get what it was, and these people evangelized it. Hewlett-Packard got it. One of my classmates from Harvard Business School worked in the group that was developing a personal computer there, and they read Ben Rosen's write-up, and Hewlett-Packard licensed it and did their own implementation based on our software. Livingston: What were the biggest conceptual hurdles for you as you were building the product? Bricklin: The original vision was of an electronic blackboard or work area. In fact, initially I also thought of it as a head-up display (like in a fighter plane) where--using a mouse together with a key pad, like a calculator with a mouse ball on the bottom or something--you could lay things out and you could use it real time while looking at people or something. So this electronic blackboard type of thing, like the typesetting layout software that was being worked on at the time. The Harris 2200 was one that I was very interested in, which nobody knows about, but I have the Seybold write-up of it. I had seen what we now call desktop publishing, because in computerized typesetting, that's what they were doing for display ads. Classified ads are automatically laid out, more or less, but in the display ads, where you're putting "Sale!" and all this stuff, that general-purpose layout--that was the hot thing, developing that two-dimensional, general-purpose layout stuff like PageMaker. The PageMaker people came out of computerized typesetting--out of Atex, which is a local company that did computerized typesetting and one of my competitors when I worked at DEC. So I had this idea, this general two-dimensional layout, and I had the idea of calculating and then recalculating, because it's like word wrap; it does that. So those ideas came up right away for me. But then how do I really express that? What exactly are the keystrokes that you do? What exactly is the metaphor? How do I make it easy to learn? I had struggled with this in the word processing world, when we invented things for word processing, because when we did Dan Bricklin 79 word processing at DEC in the mid-'70s, there weren't many screen-based word processors. A lot of them were page-based, which meant that you edited one page at a time, and if something was more than a page, you had to cut it and paste it onto the beginning of another page, because they were thinking like paper. In fact, some of them had things like platens to turn to make the paper go up and down, and you set the margins with something you slid back and forth. That was the Lexitron. But some of them, like NBI's (Nothing But Initials) system, were document-oriented. This was before Wang did their first screen-based word processor. I came out of the Multics project, which used the Runoff system, which Jerry Saltzer had developed for the CTSS (the Compatible Time Sharing System), which was one of the first time-sharing systems. To write his thesis, Professor Saltzer invented this thing called Runoff, which was used basically to do the word processing for it. It was a document-oriented word processor, as opposed to the page-oriented ones. The big word processors were the Mag Tape and then the Mag Card Selectric, from IBM. Those were relatively early in word processing. There were a few things before that, none of them screen-based. The idea of a long document that's automatically broken up and that embeds commands was like typesetting. So put those two together and we had to invent the ruler--the embedded ruler. Now, others invented it simultaneously, but we had to invent our idea of the embedded ruler that, when you put the cursor above it, it does one thing, and below it, another. In the word processors of the day, the ruler was active as you were typing and applied to what you were typing, but it wasn't really remembered in it. So we had to figure this out. We were selling it to places where secretaries would use it. People were paid by the keystroke in typesetting, in some cases. And in word processing, they were paid by the hour, which is basically by the keystroke. So we were very much into keystroke minimization. How many keystrokes does it take to do things? Hours of arguments and design about that in the typesetting world and the word processing world. I applied that to the spreadsheet. My whole mindset was, "How do I make it easy to learn to use? How do I make it minimum keystrokes for everything? How do I make it natural, so, if you're doing this repeatedly, it's the natural thing to do?" Day one I wasn't thinking computer-like. The whole idea was not to think computer-like. We used decimal arithmetic so it would act just like a calculator. We didn't use binary arithmetic, which might end up with some anomalies that you might not understand. I had Professor Jackson at the business school, and I had her look at the prototypes as we were doing it (she consulted to CEOs of big companies). She said, "You're competing against the back of the envelope. It's got to be really easy to use." I was constantly worrying about those things, and that affected the design quite a bit, because I had a lot of experience in that user interface world. I had also trained people on my product, so I had a lot of experience training people. So I knew what it was like, what people learn to use, etc. The challenge was, how do you express the value you're typing in, the formula you want to calculate, its location, and the precision of the decimal points, 80 Founders at Work and how wide are the columns and all this stuff? Is it an integer, is it a floating point number? How do you specify all that? In computerdom in those days, that was the most yicky stuff of any computer language--the format statement in Fortran, and COBOL's pictures and all that. It was just such a mess. How do you get the output specification of how it looks? I ended up with WYSIWYG, like people had done in typesetting. How do you marry that with calculating? There I came up with use of the grid as a way to be able to name things. The big problem for me was, how do you name things? How do you name the value? In the old days, it's like, variable name equals expression, right? That's how computers work. Well, this was, "What's the variable name going to be?" Today it seems so natural: you use A1. Well, first of all, it was A1, not 1 comma 1. It is too many keystrokes, it's not normal for people, and there's a whole lot of problems with it. By going to the map coordinate type of thing-- A1, G7, or something like that--that was something I knew regular people would understand. But it also parsed well: anything that starts with a letter was obviously a variable name, because numbers always start with a number, or a plus and a minus or something. So it made it really easy to make it obvious what you were typing in. So, if you said 1 + A1, I knew exactly what it was. But, if I said 1 + 1,1? So coming up with that idea, coming up with the fact that you'd be editing the output as the input--you'd basically be inputting into the output; what you see is what you get--with a separate location that showed the contents and all the attributes of it at the top, with the menu tree being shown at the top. We had very little memory space to give you in the way of help, but if you hit /, it listed all the letters you could type. If you typed a letter, it would give you the name of the command that you were doing and any options. So basically, it was always prompting you with what you could do next, once you learned to do the / key. And, of course, we could use /, because / is an infix operator, not a prefix operator, so you always knew that if something started with a /, it had to be a command. But, if it started with a +, it's going to be a number. So it was one of the few characters around that was good for that. And it was not shifted (I hate holding control keys down), and computers had used / as commands before, so it was a natural thing to use, for me. So working out those problems was the thing. But then, after that, everything else was just, "What are the required features?" Adding replication, the ability to copy a cell with absolute and relative, that was sort of a natural thing for me to come up with, and it was not uncommon in other financial forecasting systems that existed--the time-sharing systems that were not as interactive. So that just all flowed. And it was just, "What can we throw out to make this thing useful and to fit in memory?" Livingston: What kind of interesting features did it launch with? Any that you wish you had included? Bricklin: Well, it would have been nice to have a better help system, but there was no space to store that. Dan Bricklin 81 Livingston: Space was an issue? Bricklin: Oh God, the whole system, the operating system, the screen buffer, the program, and the data that you're running on, fit in 32K. A screenshot of VisiCalc doesn't fit in 32K nowadays. The Apple II only had 48K max. So this program launched in 48K, and you're going to put in a help? When Lotus launched with their help system, it was a separate disk that you put in that had the whole system in it. Livingston: Didn't Lotus design 1-2-3 with the IBM in mind? Bricklin: It ran in 256K or maybe 128--I don't remember--but you had to have an extra disk in the drive if you wanted help, as I recall. Just think about it: if a help screen has a thousand characters and you're going to have 10 help screens, there goes 10K! Where are you going to fit it when you only have 20K of memory for the whole sheet? How much are you willing to give? So I printed a reference card up, which my father actually helped me do--my father's printing business typeset and printed that whole thing for us. An awful lot of people learned the product from the reference card. It had the ability to lock--because, remember, you had a very small screen, 40 characters by 24/25 lines on the Apple II--it allowed you to lock columns or rows on the screen. They call them panes now, I think, in Excel; you could lock the panes. We called them titles. You could lock the title area, and, as you scrolled, they'd synchronize, so that if you scrolled sideways, the stuff stayed in place. It had two windows--you could actually split the window and watch two parts of the screen at once--so you could type numbers in one place and look at the sum somewhere else. And you could scroll them in unison. You could lock them in synchronous, so as you scrolled one, the other scrolled, and in one of them you might have the titles locked. And in fact, you got different column widths in different ones. Bob put in all sorts of cool stuff. They don't do that stuff today. But it didn't have commas in numbers, because we had some bugs in that. We never shipped that, which was a real problem. And all the columns were the same width. You could change it, but they were all the same width, and that was bad. If you had a label that was longer than a column and there was a blank cell next to it, it didn't automatically go into it. You had to cut it into two pieces. Those were real killers. Those were things that 1-2-3 had, among others. When 1-2-3 came out, those were the things people asked. "Does it have commas in numbers? And dollar signs before the numbers?" I think we had dollar format, which meant .00. But did it do commas, did it have variable column widths, and did it have the long labels? I remember Vern Raburn telling me those were three of the main questions that he was asked, and then people said, "Fine, I'll buy it." So those were features we didn't have that would have been nice if we did. We knew we needed those, but there was just a limit to what we could get done and would actually fit and work in the original product. Livingston: You publicly announced VisiCalc in June. When did you first ship? 82 Founders at Work Bricklin: We worked out of Bob's attic until around when we got delivery of this computer that we bought, a time-sharing computer. Bob wrote an assembler and linker for it, and I wrote an editor for it so that we could do our work. We hired an employee or two, and they helped us finish the actual product and then convert it to other machines. Bob wrote most of the code, and then this person we hired, Steve Lawrence, and myself wrote the rest of the code. I got the transcendental functions to work, the sine and cosine, stuff like that. There were bugs in divide, and Steve got those things working. We had the beta version of it ready, I think, in the late summer, together with a self-running demo version of it that was actually macro-driven--that basically had a long macro that would just run that was just keystrokes driving the thing. You could just put that disk in--the computer store could do that--and people would just leave it in the window, and it would run through an entire demo, explaining what the thing was. Personal Software sent those to every known computer store. Some of them had no idea what to do with them and just sold the demo. Some lost it. And some figured out what it was, and became rich, hopefully. In the fall of 1979, the manual was finished, production was finished, and it shipped. I think I got my first copy Saturday, October 20. Livingston: Were there any panic moments before October 20? Any times when you thought, "We can't pull this off?" Bricklin: There were panic moments in the business, but they had nothing to do with programming. We were working in a basement in Central Square. We were next to the T (the subway) right near the Kendall Square Station. The T went right by us, and every time it went by, everything would shake, because, literally, it was a few feet in front of us. We were below street level, so, when it rained, the toilets would back up. When it rained, whenever you left the building, you had to remember to turn off the toilet, or else they would back up. We missed one, and it started flooding, and the water started pouring toward our computer. I have some pictures of me there with one of those wet vacs as the water just missed our computer! Our life savings are in this one computer--life savings plus some money from relatives, plus personal guarantees on the loan. There was getting the contract finished. Dan Fylstra came over with the latest version of the contract. We didn't have word processing. We had a correcting Selectric that I was writing stuff on. Dan didn't have a real word processor or a good printer for it, but he was doing advertising, so he went to Typotech, which was a place in Harvard Square where you could do your own typesetting by the hour. So he used it as a word processor, and he would typeset the contract. Then we would be sitting there negotiating some of the stuff, and he would run off to Typotech and make changes, and he'd literally cut and paste the results. The final contract we signed--because it was up until late at night, making some changes about advances and royalties and future versions, I don't know, Dan Bricklin 83 whatever we were doing--we needed a copy of it, and it was late at night; there were no all-night things. Bob had a copier. In the old days, Xerox's patents hadn't run out, and people didn't have Xerox machines at home. We had a thing which had a lightbulb in the bottom and used this heat-sensitive paper or something, and you put one page over another, and you end up with this brown-onbrown output. And that was the actual contract that we signed. Dan had the contract in hand, and he jumped on an airplane and flew off to New Orleans, where Ben Rosen was having his conference (later it became Esther Dyson's conference), and that's where he first showed it off to people. Ben had seen a prototype, and it was being announced, semi-publicly, at the conference. So that last-minute getting that done, that was the type of thing we were doing. I wrote an accounting system. Not only did I write the editor, I wrote an accounting system for us, and I did all the bookkeeping. I mean, here I am, a business school student taught to do accounting by a wonderful professor of accounting and then taught cost accounting by Jim Cash, who's now on the Board of Microsoft, but I'm now trying to figure out doing debits and credits by hand. I didn't know what the real world had on that. And I was doing my own bookkeeping. I wrote a system to do it. Livingston: Did you have any competitors? Bricklin: We were nervous that competitors would come out. But there was just so much optimism in those days. And we were doing this as a stepping stone to do other things. We didn't know this was going to be such a big thing. We figured we'd just keep on figuring out all sorts of cool stuff. Livingston: Do you remember when you finally thought, "OK, this is a big deal?" Bricklin: It felt like a really big deal when I started having people I didn't know in the regular part of the world knowing about spreadsheets and taking them for granted. When the Wall Street Journal ran an editorial about the budget in Washington and said, "Yellow ledger pads and VisiCalc spreadsheets all over Washington are trying to figure this out," that really hit me. IBM came to us wanting VisiCalc on the IBM PC, and when they ran the advertisements on TV, they showed VisiCalc (or they showed what they said was VisiCalc; it was a mockup that they did) with Charlie Chaplin pushing a button. When Apple ran an ad, they had Dick Cavett--who had never done ads on TV before--and he would push a button and up would come VisiCalc on the screen. He didn't know what the hell he was doing, I'm sure, but I thought, "Wow! That was really cool! Dick Cavett!" One thing that really hit home was when I was going back to the airport from a conference where Ross Perot had spoken--he was the head of EDS. A few of us from Software Arts shared a limousine with senior EDS people, and they knew about VisiCalc. This is EDS, which is the big mainframe company. They said, "Oh yeah, we did some deal, and we used VisiCalc to do all the calculations for the deal." Now, here's EDS, that has infinite computing power 84 Founders at Work available, with any software of the big financial forecasting systems, and all that, and they're using VisiCalc to price multi-million-dollar deals! I found out that investment bankers who were doing real deals were using it. When the people you looked up to as the pros have switched to your stuff, that meant something. And the other was when I heard from Don Estridge, who was the head of the IBM PC project. Don had told me that, when he was about to demonstrate VisiCalc to one of the real senior people, the executive said, "No, I know how to do it. Hold on. Let me do it." And I think he was demoing on an Apple II. "Whoa!" Then you realize that you did make a mark, and people did get it. Livingston: VisiCorp and Software Arts had some legal disputes. Is there anything important that you learned from that? Bricklin: Stay out of lawsuits if you can help it. It's bad for both sides, especially small businesses. That's lawyers' business, to them, solving things through lawsuits. But it's very, very expensive. It's a sport of kings, and it uses up a lot of time. Unless you're a very big business that can make it a very small part of what you do, it's much better to find other ways to solve things. Frequently, individuals can do it better face to face. People who are the heads of companies understand that. The boards involved there let it happen, and they shouldn't have, since it ended up being bad for both companies. Livingston: And it distracted you. Bricklin: Distracted? It killed us. We had just finished negotiating a deal to sell our company, for cash--a lot of cash--to a major company. It would have changed the whole industry. We had been approached by H&R Block to buy our company for, I think, $50 million in cash, plus stock. It was based on the numbers we had. It was kind of bogus, but whatever. They had a division called CompuServe, and we were going to be bought by CompuServe. We had board approval from both sides. We got sued a day or two before the deal was consummated. This was not very good. I was used to bad things happening at the last moment. If that had happened, we would have ended up with all the stuff we were doing over at CompuServe. The world would have been quite different. One of the pioneers of the Internet, David Reed, worked for us. He would have worked at CompuServe instead of at Lotus, because it ended up, when things went down, Lotus bought us out. Thank you very much, Lotus! It was the right thing for them to do, business-wise. But also it was the right thing for them to do, and Mitch [Kapor] was very good about that, to save us from bankruptcy. It was just a few million bucks to take us out of our misery, to pay off our loans. But we weren't able to run the business. It killed the deal; we weren't able to sell the business while we were in a lawsuit. VisiCorp was in bad shape. Their legal fees were running about the losses they had every month. It killed VisiCalc--well, VisiCalc was being killed by 1-2-3 anyway. They thought the new product, VisiOn, would have saved the day, but new products don't do very Dan Bricklin 85 well right away, often. It was a precursor to Windows in the days when the PCs weren't powerful enough to do it. So for all of its advancing the art of things and cool stuff they did, it wasn't its time, and they ended up selling it off to make some money, and they ended up going belly up. It was bad all around. What I do realize is there are advantages to selling at a peak. You don't know when the peak is. I know people who sold their businesses when everybody thought you were crazy. "The business is going through the roof; why are you selling now?" And in hindsight, of course, it turned around. Six months, a year later, the business started crashing. They didn't get the peak, but they came pretty close. There are some people to whom it's worth taking the risk, because you risk going for the big one, and, in a portfolio, that's good. But as they say on Wall Street, the bulls make money, the bears make money, but the pigs get slaughtered. In other words, don't be greedy. Whether you think things are going up or things are going down, you can make money going both ways. But, if you are piggish, are greedy, that's when you have problems; you'll be irrational about that. It is worth it sometimes, if you can do it, to reach for the stars. Microsoft didn't reach for the stars. Microsoft was step by step by step to where they got, and it was profitable all the way to it. So that's the traditional way of doing it. The Google, Netscape way, those things, sometimes it works, and sometimes-- usually--it doesn't. But sometimes it does, and the payoffs are incredible. But, if you're a business person who wants your business to succeed, as a business, because you like that business, you take a different view. So the risk profiles are different. A lot of people make money because they're very good at timing. We were close. If we hadn't been sued, I would have done pretty well financially, because Bob and I owned most of the company at the time. And we would have had an interesting next step, going into what was probably the leading online business at the time. Maybe it would have ended up into the Internet or something, or it would have made the jump better. Who knows? But it didn't happen. Livingston: Do you have any regrets? That one was out of your hands. Bricklin: Yeah, it was out of our hands. If we had been able to settle in advance, the thing would have closed, and we would have made a lot of money, and we'd have a bigger house, and whatever. But, you know, as I always tell people, here it is, 25 years later, and you're still interviewing me. There's fame and fortune. I didn't get much fortune out of it, but, on the other hand, the fame has basically given me a meal ticket ever since, and I learned a lot from it, and the rest of my life has been pretty good. All in all, I can't complain. I did a lot better than I ever expected to, in all sorts of ways. So there are no regrets about that. I mean, each thing, you think, "Well, if I had put this feature in, it would have been better." Livingston: Do you remember any disagreements that you and Bob had? 86 Founders at Work Bricklin: Oh, we had lots of disagreements, all the time. People always thought the company was going to die, because we'd yell at each other on all sorts of issues. It was usually over technical stuff. Bob's much more aggressive in many ways than I am, and I'm much more conservative. So we're very complementary. While I'm disorganized, he's more disorganized, in certain things, so he depends on me for the drive to get things to completion. On the other hand, I depend on him for some of the reaching for the stars. So we were very complementary, but that's tough. It's like having old married couples who spat all the time, always yelling at each other. It wasn't as bad as some businesses, where it actually is a married couple. But our friendship has continued to this day. As people know, in the business--like Bill Gates is known for this, about being really tough in meetings, and arguing and stuff like that--that's just a way of testing your own understanding of things. By arguing with others about it, that's how you learn. And, if somebody can't take the arguing with it, then maybe they don't really believe in what they're talking about and they don't understand it well enough. We'd argue and then we'd go out to lunch together, because it wasn't based on animosity. We had enough problems with people outside. Livingston: Do you remember a time someone tried to take advantage of you or cheat you? Bricklin: We needed to move, so we bought a building and rehabbed it, because it was not in the best of shape. It was an old factory, and we turned it into programmer heaven. It turned out that we spent too much time on that, and we should have spent more time on the product. So stick to the knitting, and focus. But, we did really well by that, and, basically, the only money I got-- other than my salary--out of Software Arts was the money I made on selling the building. When we got the building, we got a loan to pay for it. We had a bank we'd been working with for years at the time, and we told them we wanted to do a loan, but we wanted no personal guarantees. When you have personal guarantees, they'll take your house. So, we wanted no personal guarantees. And the bank said, "Sure." We came down to the last closing papers, and we looked at the papers, and what does it say? Personal guarantees! They said, "Oh well, that's standard. We always do that." We got another bank, and sure enough as they were about to close, in came the personal guarantees. It wasn't until the third bank--we finally got one--that we did it with no personal guarantees. Livingston: How did Lotus end up buying you? Bricklin: At the last minute, when the company was about to go under, we found some people who were willing to buy the company, but they wanted me to spend a year working for them, and I was not happy about this at all. I ran into Mitch Kapor on an airplane, and we talked. That's Monday. Friday night, Lotus bought our company--they bought the assets of the company. So finally we sold the company, and I'm out, with no strings attached. That was great. And Dan Bricklin 87 then we have to finally sell off all our stuff, because they did an asset deal. So people had to stay on to close down the company and all the liabilities for a year or two; it was a mess. All those things happen, all the time--the wonderful ups and horrible downs, but that's what business is all about. And it's very personal. There are a lot of personal things. It's running into people. And how did I know that I should talk to Mitch? Well, our insurance agent was also his business insurance agent, and he talked to Mitch. So I knew that Mitch knew what was happening with our business. He knew our business was in rotten shape, because we had to work out all this stuff about insurance and getting the right things done and what we were going to do in case we were going to declare bankruptcy, because we came within days, within minutes, of declaring bankruptcy at one point. And so we had to work closely with him. But, he also worked closely with Lotus. So he was able to tell me, "Look, why don't you talk to Mitch? He's a good guy. He'll help." I had known Mitch from the Apple II days and all those other things. And we were like sister companies until they competed with us, and even then. And a lot of our people were at Lotus, so I liked a lot of the people there. So when I ran into Mitch, I was willing to actually tell him how bad things were. Even though it seems like it's big business and impersonal, and "they" take care of it, it really isn't. There is no "they." It always comes down to an "I" of somebody, and in many cases, it's a principal.