[btilly]

No, it isn't. A lot of companies that sell software now sold other things before they sold software.

Many companies move into software, but of the companies listed on that Wikipedia page, only IBM, HP, Lockheed Martin and arguably Accenture did something else before going into the software business. (Accenture was spun out from an accounting firm, so in some sense they did accounting first, and in some sense they existed long before they actually were a company.)

And of the companies listed on that wikipedia, page, only IBM, EDS, Lockheed Martin, CSC, Capgemini and SAP existed before the 1980s.

Check again. Microsoft was founded in 1975. Oracle in 1977. And while Accenture did not exist as a separate company before 1989, it started as a department at Arthur Andersen that had been selling software consulting since 1953.

Nearly all of those did other things before entering software.

What exactly did EDS, Computer Sciences Corporation, Capgemini and SAP do before entering software? As far as I can tell, nothing.

In any case, the exception makes the rule: virtually nobody was a professional coder before the early 1980s. Of the few who were, many retired rich and young after two of the largest technology booms in world history. They don't make good examples of career longevity, for what should be obvious reasons.

Again false. There were a lot of professional coders before the early 1980s. Most worked in the mainframe world. Much of their code is still running today. And very, very few of them retired rich and young. (Certainly the ones that I know didn't.)

[timr]

"Check again. Microsoft was founded in 1975. Oracle in 1977."

Fine. The software industry is 35 years old. You're picking nits.

"What exactly did EDS, Computer Sciences Corporation, Capgemini and SAP do before entering software? As far as I can tell, nothing."

The same thing as that Arthur Andersen group that you're classifying as part of the "software" industry. Mainframe hardware sales and consulting:

"GE asked Arthur Andersen to automate payroll processing and manufacturing at GE's Appliance Park facility near Louisville, Kentucky. Arthur Andersen recommended installation of a UNIVAC I computer and printer, which resulted in the first commercially owned computer installation in the United States in 1954."

I never said that there weren't any software developers before 1980 -- I said that there wasn't a software industry, as we know it. That mainframes were first sold in the 1950s is not evidence to the contrary.

[btilly]

I'm very emphatically not picking nits here. See http://www.softwarehistory.org/history/ec.html for an overview of the history. From there you can dive into the early history of the software industry, with names, companies, and more. (They have less detail than I'd like.) For example if you read http://www.softwarehistory.org/history/csc.html you'll find that Computer Sciences Corporation started with $100 and a contract from Honeywell to write a programming language called FACT. They were NEVER in the mainframe hardware sales business. As for consulting, guilty as charged. There was no idea of a market in software, so if your company wanted to get paid for writing software then consulting was your only chance. But that still happens in the software industry.

However despite there not being a recognized market, software that was produced for one client could and did get sold to others. And as the 60s progressed there came to be enough of these pieces of software out there that the first regular catalog of software products for sale got going early in 1967. By the end of the 60s there were hundreds of software packages for sale, and IBM declared that it would get into that business starting January 1, 1970 by unbundling some of its software from its hardware.

By that point there is no question that there was such a thing as a software industry. Small? Yes. But definitely present at a surprisingly early point.