PC software: the first decade

by Evan Koblentz

NEEDHAM, MASS. -- Anecdotes were shared, history was clarified, and lessons were learned, all firsthand, at the PC Software: The First Decade conference here last week.

"We got started to some extent because Fortune magazine said Bill Gates invented the software industry. He didn't," explained Burt Grad, president and co-founder of the Software History Center (www.softwarehistory.org/home/PCsoftware.html), the Westport, Conn. conference sponsor. But the industry's pioneers, several of whom attended this year's conference, can partially blame themselves for the mainstream media's misunderstandings: "We were all too busy doing... we didn't record a darn thing," noted Larry Welke, a trustee of the Los Altos, Calif.-based Charles Babbage Foundation.

VisiCalc -- the spreadsheet program that made the term "killer app" synonymous with the microcomputer revolution -- illustrates those points. Decisions such as which hardware to support, what to name the product, and whether developer Software Arts and publisher Personal Software (later VisiCorp) should merge or sue each other are all slightly different stories today, depending on who tells them.

In general, Apple was the best hardware choice because of its openness and the quality if its disk drives, agreed VisiCalc co-inventors Dan Bricklin and Bob Frankston, along with Personal Software/VisiCorp co-founder Dan Fylstra. But the fact that Apple keyboards only had two arrow keys was a problem, Bricklin said. Years later, besides the Software Arts (Bricklin/Frankston) and VisiCorp merger that never happened, there were other close-but-no-cigar events along the way, Bricklin said, such as a deal with H&R Block/Compuserve that came within 48 hours of closing. So if either the original SA/PS merger or the Compuserve deal had transpired, then today's software industry might be a vastly different landscape, the founders agreed. Mitch Kapor's leaving the VisiCalc team in favor of co-founding Lotus Development Corp. might also have been affected, for example.

The good news is that Bricklin and Fylstra, sparked by the conference, each re-examined the whole story in new papers that will likely soon be made public. (We'll publish hyperlinks to these papers when they are finalized, which should occur within a month or two.)

Although it's been 25 years, it is good for history -- and provides valuable insight for collectors -- that the whole story of this groundbreaking software is finally becoming known. (Equally exciting is that Bricklin recently left his CTO post at web hoster Interland Inc., and is beginning to develop new software again. For his direct explanation, see www.softwaregarden.com/about.html.)

I also attended conference sessions on development, testing, and maintenance, and on customer support, documentation, and training. In the former, alpha and beta testing of PC software in its early years was "purely a gimmick... in most cases they were disruptive in giving you bugs," said Mike Maples, a former software executive at IBM and then at Microsoft Corp. In the latter, the computer book industry may have been partly responsible for software piracy in the 1980s and 1990s, because of how easy they made it for consumers to get the equivalent of a product manual without actually buying the software, well-known writer and trainer Adam Green said.

Looking forward, software's history will be recorded through the new Information Technology Corporate Histories Project, a Sloan Foundation joint venture between the Charles Babbage Foundation, Computer History Museum (of Mountain View, Calif.), and Software History Center. Staff members will first study the timesharing industry, and later will explore industries such as databases, financial services, mainframe and PC applications, and storage, project leader Luanne Johnson said. The research will include oral and written recollections, via online and in-person methods, she said. "We're going to need hundreds of volunteers," she added. She can be reached at luannej@sprynet.com.

Overall, "We felt the meeting was a real success... We captured a lot of good material," said Grad, of the Software History Center. "Their openness and willingness to share was terrific," even among industry figures who've had differences in the past, he said. Transcripts will probably be available in about six months, either online or in print, he said. The previous conference two years ago focused on mainframes and this year's was for microcomputer applications, so the next one in 2006 will focus on minicomputers and small-business systems, he said.