Barry MacKichan
CEO of MacKichan Software, Inc.
Barry MacKichan is an entrepreneur, software developer, and mathematician. He holds an AB degree from Harvard and a Ph.D. in mathematics from Stanford University. After an academic career that included the Institute for Advanced Study, MIT, Duke, and New Mexico State University, he got hooked on writing software while writing a word processor to help writing mathematical papers.
He co-founded TCI Software Research, Inc., in 1981 and was CEO until 1988, when he left to join Microsoft. There he worked on Microsoft Mail, the Microsoft Foundation Classes, and OLE 2.0. He retired in 1994. In the meantime, TCI Software Research was bought by International Thomson Publishing; the acquisition was not successful, and so MacKichan bought back the assets of TCI Software Research in 1998. Now under the name MacKichan Software, Inc., the original owners and developers are proving that specialty software products need to be developed and marketed by small, nimble, independent companies.
MacKichan lives on Bainbridge Island, Washington, with his wife Lynda, who is the business manager of MacKichan Software. They have four grown, accomplished children, and enjoy raising llamas in their spare time. Barry also does large format and wildlife photography when time permits.
Resonance: In 1981 you co-founded TCI Software Research. What need did you see unfulfilled at the time?
MacKichan: The state of the art then for writing mathematics papers was the IBM Selectric typewriter with a drawer full of type balls. Word processors had arrived for most people, but for mathematicians the fastest way to write mathematics still was with a pen and paper.
Several of us had written a word processor for Terak microcomputers, which were the first ones we knew of that had re-definable character sets. When the IBM PC came out, I read the assembly language listings in the technical reference manual and saw that we could display whatever symbols we wanted in graphics mode. That was the key. We started our company to write and market a technical word processor for mathematicians, called T3 , which supported our company for the next ten years.
Resonance: What are the trends you see in software development?
MacKichan: The sizes of the chunks that programmers work with will continue to grow. We already have gone from single machine instructions (assembly language) to gluing ActiveX objects together. Software systems will more and more be built by linking larger and larger objects together. The use of web browsers as a generic user interface is an example of this.
Resonance: How do you see the Internet affecting the way we view information and communications? Is the Internet an invention of the same magnitude as the printing press?
MacKichan: The Internet will have a dramatic effect, but I don’t believe it will be as dramatic as the effect of the printing press. They are both quantum jumps in our ability to communicate, but necessarily the first and second jumps (the invention of writing and the printing press) have a more dramatic impact than the later ones.
The Internet makes information, which was available before, available at less cost and greater speed. The change will be more incremental than revolutionary. The relative costs of various types of information will change. This doesn’t mean that the change won’t be dramatic: many of the ways we do business will change, and there are quite a few occupations that won’t last through the transition. The change in politics may be as great as the change wrought by television – which has been dramatic.
Resonance: Following up on your last answer, what occupations do you see disappearing, or changing so dramatically that they are in effect new? Also, in politics, do you see the Internet becoming the main medium of election campaigns, perhaps altering the way campaigns are financed?
MacKichan: A lot of occupations exist because information is hard to find, and the web potentially makes it easy to find, which threatens those occupations. Almost all intermediaries fall into this category. I would be worried if I were a software or book distributor, a real estate agent, or even, in the long term, a car salesman.
In some cases we will continue to pay, willingly, for face-to-face contact. Teachers will still be important to our children, for example. In other cases, we’ll find we never wanted the face-to-face contact, at the Motor Vehicle Department, for example.
The Internet will be important, it already is, in politics and fund raising. I think it will become the main medium only when it is integrated with television. A downside risk of the Internet is that it strengthens fringe groups and single-issue groups by making it easier for like minds to find each other to the exclusion of all others. I worry about the effect this might have. After all, the real purpose of the political system is not to impose your view, but to reach consensus, and anything that retards that can be a step backwards.
Resonance: Has the Web created synergies that could otherwise never have existed?
MacKichan: Yes, just as the automobile, airplane, and telephone did.
Resonance: Where do you see computational power heading in the next decade? Are computers of the type depicted in science fiction, such as Star Trek, a near-term reality?
MacKichan: I’m just an observer in this area. It seems like the exponential increase in memory and processing power will continue for quite a while longer.
When I was at Microsoft, I tried to get people to discuss how to best use 2000 MIPS – to think about how it would change what we can and should do with computers, and how the user interface should change. I never did succeed in getting a continuing discussion going. Now 2000 MIPS looks pretty modest.
Resonance: In particular, we are starting to see voice recognition software becoming available for word processing applications. Is this the beginning of the end for the keyboard in the modern office?
MacKichan: We have just announced voice input for our products, which was developed by another small company, Metroplex Voice Computing. It is pretty awesome to say “3D wave equation” and to see it appear on the screen. Still, once you master it, the keyboard is pretty efficient and probably will remain a component of almost all computers for a long time. I see voice as another way of increasing the bandwidth between the human and the computer. I’ve found that mixing voice and the keyboard is faster than either one alone: I use voice as the way to call macros, and the keyboard to enter plain text.
A laudable use of voice recognition is making the computer and its software available to users with disabilities. Now it is possible to speak to one of our programs to display mathematics, to solve equations or to simplify expressions. And the output can be in Braille. The only piece missing is voice output for mathematics, and the MathML working group has addressed this problem, so I expect it to happen in the next couple of years.
Resonance: You were a member of The Institute for Advanced Study in 1968-1969 and then 1975-1976. What was the focus of your work at that time?
MacKichan: The first year I was pursuing research questions that came out of my Ph.D. thesis on over-determined systems of partial differential equations. The second session I was working on problems in the theory of several complex variables relating geometry to sets of equations induced on submanifolds. These sets of equations are over-determined systems of partial differential equations, so my field of research hadn’t really changed. A common thread in my interests was the relationship between geometry and analysis.
Resonance: Could you give us your impressions of the differences between the academic and entrepreneurial worlds?
MacKichan: There are very smart people in both worlds. One of the attractions of Microsoft was that the people there reminded me of a good graduate school.
What I like about the entrepreneurial world is that the goal is really very simple: to produce something that a lot of people really want or need and will pay for. Even if 99% of the people don’t want it or care for it, the remaining one percent can make you succeed.
In academia, the market is very much smaller. You can spend several years writing some papers, and the group that will judge whether they are good or great is frequently much smaller, perhaps a dozen, and in such a small group, random factors such as personal rivalries and personality quirks can make quite a difference. This is the reason, I think, why politics plays such a role in academics. I’d rather my products be judged by a larger market.
Resonance: A key product of MacKichan Software is Scientific Workplace. I have used it often and find it indispensable to my scientific word processing. One feature that I especially appreciate is the embedded subset of MAPLE. Where do you see this package heading in future incarnations? Do you see embedding a basic drawing tool so that authors can easily create publication quality schematics and sketches for their manuscripts?
MacKichan: One of the things I worked on at Microsoft was OLE 2.0 (Object Linking and Embedding), a mechanism for embedding objects in documents. In this case, the program that manages the document doesn’t need to understand the data format of the embedded object. It simply hands the data to the server, which modifies the data and presents its own interface to the user. An embedded drawing tool would fall into this category.
Our integration with Maple is different in that we create the data – an equation or a mathematical expression – without invoking Maple, and then we translate it to send it to Maple. The link between the applications is not at all like OLE, since it based on a flow of data that is manipulated by both applications. When Maple has done its work, we translate Maple’s output to our form – a subset of LaTex, a typesetting language for mathematics. This translation from LaTex to Maple is not trivial. You need to apply heuristics to decide if f(x+y) is a function applied to the quantity (x+y) or is the product of f with the quantity (x+y), and to know that dy/dx is not a fraction. You have to decide the same things to know how to do voice output of mathematics, by the way. The heart of the problem is that LaTex contains information on how a mathematical expression appears, but it does not include the mathematical content. It is indifferent to whether dy/dx is a derivative or a fraction.
Our products currently don’t support OLE 2.0, but this will change. At that time you will be able to use any embeddable drawing tool; I expect that we will not be writing one, since our time is better spent on our special area of expertise, providing a great interface for interacting with mathematics.
Integration of the type we have with Maple is improved by standards for the data that is passed between the two applications. There is currently an emerging standard for the representation of mathematics called MathML (Mathematics Markup Language). We are on the MathML committee, and I expect that we will support MathML in the near future. MathML is an application of XML (eXtensible Markup Language), which allows tagging all sorts of data, which will encourage the sharing of data between applications. This is a big step forward, I believe.
The technology we have can be applied to the problem of entering MathML. Users, we think, will continue to want to type simply dy/dx or f(x+y) without having to tell a program that they are writing a derivative or a function. We expect to be able to use the same heuristics we are using with our Maple connection to make the entry of MathML as simple as our current interface.
Resonance: Can you tell us about other project(s) you are currently working on?
MacKichan: Running MacKichan Software is taking almost all my time. Before I bought back into business, I spent some time working on literate programming tools. These are tools, first developed by Donald Knuth that allow you to write a program as you would a book: you can write paragraphs of design and explanation interspersed with bite-sized fragments of code. The tools will then produce a nicely formatted document explaining the program and compilable code that produces the program. I was modifying the noweb system to produce XML output, so that the document explaining the program would be a fully cross-referenced web page. The intention was to produce tools that would integrate with Microsoft’s Visual Studio, but now the project is on hold.
Resonance: Looking at the kind of authoring tools that exist, both for scientific as well as other writing, and marrying them with the revolutionary Internet, what do you think is the future of book publishing?
MacKichan: I’ll get out on a limb and predict that books will not be obsolete soon, certainly not in my lifetime. The technology of a finely bound, carefully printed book is superb.
Having said that, I also think that most of what is printed now will be delivered online in the future. Reference information, especially information that is soon outdated, shouldn’t be printed.
The eventual outcome will probably depend on the type of book. For the first type, a self-contained book such as a mystery novel, the issue will be settled by economics and ergonomics: bits are cheaper to deliver, and paper currently is easier and more comfortable to read. I think that electronic delivery will eventually win except for enduring literature. Shakespeare and Melville will always be available on paper, I hope.
For the second type, a book with references, such as a history book, the value of the electronic form increases as the number of references in electronic form increases. This is sort of like Metcalfe’s Law for networks (which says that the value of a network increases as the square of the number of users). The real payoff comes in converting a library to electronic form, but the payoff of converting an individual book to electronic form is not as great. When the library is converted, the reader can link to all the references, perhaps even to the primary sources.
I think the area where the electronic form has the greatest advantage over the printed form is in the mathematical and technical areas, and this is the third type of book. You can argue (and this is not just marketing hype) that what appears in a Scientific WorkPlace document is not just a representation of mathematics — like ink on a page — but is a mathematical object. You can view it as a mathematical expression; you can also compute with it; graph it; run a simulation; or modify it. If you have a spring or an electrical circuit, you can run a simulation, modify it, and run it again. The tools for doing this are just starting to appear, but I predict that the value of electronic books and online courses will be the greatest in the technical areas, in spite of the fact that these are the slowest areas to convert to electronic form because of the problems of representing mathematics. We see our Scientific Notebook product as a first step in the direction of providing a reader for such electronic books.
Resonance: Who are your favorite writers?
MacKichan: Hmm. One that influenced our lives is Wendell Berry. His book, The Long-Legged House, which my wife and I read in the late ’60s, changed our lifestyle.
Another one in science is Stuart Kaufmann, whose books relating biological questions to complexity theory I find fascinating. The mathematics in them is not rigorous – he will assume a mathematical theorem applies to a situation in which the conditions of the theorem aren’t satisfied exactly – but the ideas are fascinating. I enjoy reading almost anything that comes from the Santa Fe Institute.
I have always loved wilderness, and so I enjoy works relating to the North American wilderness – histories by Francis Parkman, Bernard DeVoto, and David Lavender. I also have read most of what Viljhalmur Stefansson wrote about his explorations in the Canadian Arctic in the teens and ’20s- I shook his hand when I was in eighth grade. And Farley Mowat, who visited, and wrote passionately about, the Eskimos west of Hudson Bay where I took some canoe trips and a dog team trip when I was younger.
Resonance: What are some of your skills that permitted you to succeed as a mathematician, software architect, and entrepreneur?
MacKichan: I think there is a lot in common between mathematicians and software architects. My thesis advisor, Prof. Don Spencer, stressed that mathematics was about structure, not just theorems. The point of research is not to find more and more facts, but to understand how everything fits together.
Good software architecture should flow logically from a few basic, orthogonal principles, just as mathematics flows from a few axioms and definitions. When the user interface is designed this way, I think that it is seen as reasonable, predictable, and understandable even by users who aren’t mathematically inclined and who couldn’t tell you why they think the interface is reasonable, predictable, and understandable.
As an entrepreneur, I’m still influenced by my father. He started as an academic in Electrical Engineering and tried several ventures before establishing a consulting engineering company that at one time was the only company designing power transmission systems and surveying them with aerial photography. I worked in his office during the summers as I grew up and I’m sure that a lot of what I know about running a business was absorbed during those years. Coincidentally, we both made the transition from academia to business at the same age.
Resonance: What were some of your other interests?
MacKichan: I have always loved the wilderness. Starting when I was 19, I took three long canoe trips in northern Canada. The first started near Winnipeg and went to Hudson Bay. Once we got to Hudson Bay, we worked our way up about 60 miles of the Nelson River and portaged about 7 miles to the railway. The whole trip was about 750 miles. The other trips started farther north and ended at the Eskimo settlements Eskimo Point and Baker Lake.
Back then there weren’t GPS’s or EPIRB’s, and radios were heavy and bulky, so we were completely isolated from the rest of the world for most of six weeks. If something happened to us, it would be weeks before anybody started looking for us. We had our adventures: once we lost a canoe in a rapids and had to swim after it. We caught up with it a few hundred yards before the next waterfall in the river. We swamped in a rapids and gashed the canoe. On one trip we took four paddles for two people, and three broke. We learned a lot about improvising and making do.
Later, I realized that this was great training for launching a business. The situations are very similar: you plan, you prepare, you worry, but the only way to move is to make the leap, take care, use your wits, and don’t look back. More recently, I’ve indulged a love of wildlife photography (35 mm.) and large format (4″x5″) black and white photography, mostly of natural subjects.
Resonance: You have mentioned the wilderness and nature several times as being significant in your life and its path. Why is it special to you?
MacKichan: Partly it comes from my mother. When I was eight and my brother was eleven, she took us on a canoe trip in the Minnesota canoe country. It was the first time camping for all of us. Mom had always wanted to go camping, but had never had the chance and was intimidated by the popular myths of conquering the wilderness. Then she read a line in Canoe Country, by Florence and Lee Jacques, that said you aren’t allowed to take a gun into the Quetico Park. Mom concluded that therefore a woman pushing 40 with two small boys would be safe, and off we went. Dad wasn’t interested, and didn’t understand why we would want to sleep on the ground.
We went for three years. We never paddled very far, and we must have looked comical on the portages, but we had fun. Several years later, a friend and I turned the tables and started taking Mom on trips, seven or eight summers altogether. She did the cooking, and we did the paddling and portaging. This friend and his brother were with me on my first long canoe trip to Hudson Bay. I was 19, and they were 17 and 15.
Maybe because I took these trips while I was coming of age, they had a big effect on my attitudes. I learned a lot about what is essential and what is not, about self-reliance, and self-confidence. I got so I feel very comfortable in the wilderness, and a have a fondness for bleak, windswept places. It is saddening to see the remoteness of these areas dwindle away slowly.
While in graduate school, I got up to the Sierras as often as I could, as an antidote to the pressures. This was when I got interested in wildlife photography, but it was a long time before I had the time and equipment to pursue that seriously.
Resonance: What are the roles of art in society? Does art do something for society that cannot be done by anything else? How are art and science intertwined? Does science have a role in art? And what about the role of art in science?
MacKichan: For me, photography provides a counterbalance to the sometimes-excessive rationality of software development. In general, art provides an outlet for our intuitive sides that wouldn’t be acceptable in our daily professional lives.
I think art is intertwined with just about everything. So is science. As a photographer, I use technology all the time.
I’ve always thought of mathematics as more art than science. It’s true that proving a theorem is a very analytical process, but the real accomplishment is knowing what theorems are provable, and making definitions that create interesting mathematical objects. This is more art than science, and it comes from a different place than rationality.
One interesting influence of mathematics on art is that of fractals. You see the influence of the beautiful graphics of the Mandelbrot set in more and more art.
Resonance: Your training was as a mathematician; your Ph.D. in mathematics from Stanford University. What attracted you to mathematics? What was your research specialty? Do you still follow the work in this area?
MacKichan: I decided when I was in high school that I wanted to be a mathematician. I had some early successes, some expository papers published when I was in high school, that encouraged me.
My research area was over-determined systems of partial differential equations and several complex variables. (I have found that this answer almost always stops a cocktail party conversation, usually with a confession that “Math was my worst subject in school.”)
I haven’t followed work in the area since I left academics in 1982.
Resonance: Speaking to a young person of about 13-14 years old, just entering High School, what advice would you offer?
MacKichan: Learn all you can, but remember that the goal is to learn how to learn. The world is changing so fast that you will have to continue learning all your life. Also, if you have a chance for a big adventure, such as I had in canoeing in northern Canada, go for it. The lessons will stay with you for the rest of your life.
Resonance: What do you consider to be a good first programming language for a teenager to learn?
MacKichan: Java. I use C++, but it is carrying a lot of baggage from C. Java is what C++ would be if it didn’t have to remain compatible with C. It satisfies the criteria for a first language: it is object oriented, it is simple, and the design is clean.
Resonance: If you were having a discussion with any living person, who would it be and what might some of your questions be?
MacKichan: There is a man, I don’t remember his name, who joined Admiral Byrd as a dogteam driver on one of Byrd’s Antarctic expeditions. Byrd named a mountain for him. A couple of years ago, as a man in his eighties, he returned to climb the mountain.
I always love hearing about a great adventure, so I’d certainly ask him about the trip. But what really interests me is how he has kept the vital spark burning so brightly for so many years.
Resonance: What do you do in your free time?
MacKichan: My wife Lynda and I raise llamas. This is something we started before we got back into business, but we are glad for the daily ritual of taking care of them and relating to them on their terms. It brings us down to earth twice a day. And the babies are so adorable.
Resonance: If you could solve just one of this nation’s/world’s problems, which one woud it be?
MacKichan: I used to be very passionate about saving what is left of the world’s wilderness, and then I thought that we would be doing well to survive our own pollution and consumption. While I still am concerned about these, I am not as pessimistic as I used to be. Glen Canyon dam will be removed, but it might take 100,000 years for the river to grind it down. I now believe that our over-consumption of resources and pollution will stop, whether from our enlightened self-interest or because of increasing catastrophes. Enlightened self-interest still has a chance. I think that the change in attitude that this entails may bring about a renaissance in the values of art, culture, and community. This may be a slow process.
The evolution of the computer and communications may play a part in this, as communication comes to involve less physical travel and fewer resources. I think our worlds will simultaneously become more cosmopolitan and more local.