The spreadsheet was a remarkable  invention and yet the people who pioneered it didn't reap all the rewards they should have. Today we take spreadsheets for granted, but there is fascinating story lurking behind the scenes.

The strange fact is that the company that built its fortune on the spreadsheet - Lotus - didn't think up the original idea.

The first spreadsheet on the market, VisiCalc,  was invented by Dan Bricklin and programmed by Bob Frankston. They thought up the idea and implemented it, but unlike other software pioneers they chose not to market the product themselves. Instead they contracted this role to Personal Software, a company run by Dan Fylstra. Visicalc was so successful a product that eventually Personal Software changed its name to VisiCorp.

There were a number of problems with this situation. The first was that in those early days the economics of selling mass market software wasn't well understood. Fylstra's company had offered a 37% royalty to Bricklin and Frankston's company Software Arts. This is about twice too much by today's standards. VisiCorp struggled to make much money out of their most saleable product.

The second problem was that at the time it was generally believed that software wasn't patent-able. Patents generally cover ideas which are expressed as machines. A program is more like the text of a novel or a technical manual and this implies that it is more reasonable to protect it by copyright. However copyright doesn't protect the idea expressed in a book or program only the particular expression of it. What this means is that copyrighting a program stops another programmer "lifting" it line-for-line but it doesn't stop them writing a program from scratch that does the same thing.

As it turns out opinion at the time was wrong - it is possible to copyright a new idea that is expressed in the form of software. You can't patent software that does something well known, even if it does it so much better that it is considered a breakthrough. It must contain the essence of a completely new idea.

The first program was patented in 1981, which was much too late for VisiCalc which had been issued without patent protection in 1979 by Bricklin and Frankston. With hindsight it is quite obvious that the spreadsheet idea was patent-able.

There were clones of VisiCalc almost from the day it appeared but none managed even to dent its market leadership. It was first and, most importantly, it was on nearly every machine you could think of. In those days this was important because there were dozens of PC standards. To make porting to new machines as easy as possible VisiCalc was written using the UCSD P system. The P system was a low level interpreter for P code. To make it run on any machine all you had to do was write a fairly simple P code interpreter in the machine's native assembler and then all P code programs would run on it unmodified. This was a great advantage but it had one serious disadvantage - it often produced a slow program.

Mitchell Kapor born 1 November 1950

Mitch Kapor also had a marketing deal with Dan Fylstra's VisiCorp. During his time at the MIT school of management he had written a simple modelling language called TROLL. This was rewritten with a colleague to run on microcomputers as an add-on to VisiCalc. Spreadsheets are responsible for starting the "add-on" market which, courtesy of OLE and VBX, spread to all other areas of applications software. Kapor's programs were sold as VisiPlot and VisiTrend. The pair of programs earned around $500,000 over a two- or three-year period. Then in 1982 Dan Flystra offered to buy the programs outright. Kapor agreed and netted $1.2 million which made his share $600,000.

The second important person in this story is Jonathan Sachs. He worked for Data General after leaving MIT with a maths degree. While at MIT he invented a language, a derivative of Forth called STOIC.  His work at Data General was on an operating system and he didn't enjoy the management aspects, the team work or the operating system.

He left and set up a small two-man company with a fellow DG colleague. It wasn't successful but he did write a spreadsheet program for DG minicomputers. The company was dissolved and Sachs left with the rights to the spreadsheet program. He felt that much of the problem was that neither of them knew anything about marketing - and this is why he approached Mitch Kapor with the idea of developing it further.

The IBM PC had appeared at the end of 1981 and Kapor thought that it had a future. At that time it wasn't 100% certain that the world would become dominated by the IBM PC and its clones. There were lots of alternatives for the future. The PC had two front runner spreadsheets - VisiCalc and MultiPlan. The trouble was that they were both ports from other machines. VisiCalc for the PC was a conversion of the TRS-80 version, an 8-bit Z80-based machine. MultiPlan was Microsoft's Mac based spreadsheet.

Users of the new and potentially powerful 16-bit 8088 based IBM PC were disappointed in VisiCalc's performance. A lot of early PC software treated the machine as an 8-bit 8080 and even restricted users to 64KBytes of data. VisiCalc did exactly this and made no attempt to use the new features that the IBM PC had and to make things even worse it was very, very slow.

Bricklin and Frankston were so busy fighting with Dan Flystra's company about rights to market VisiCalc that they didn't see the importance of the IBM PC. It was just another machine that they had to support. By contrast Mitch Kapor was prepared to bet his whole future on the PC, and more precisely on PC-DOS.

At the time the IBM PC was shipped with two competing operating systems PC-DOS and CP/M 86. You could also add the UCSD P system as a third possibility. Mitch Kapor not only thought that the IBM PC was going to be important, he decided that PC-DOS was going to be the dominant operating system.

<ASIN:1400082471>

<ASIN:0071358951>

<ASIN:0911625658>