A reality check on 'open source'

When The Washington Post runs a front-page story on "open source," as it did in December ("The Spreading Grass-Roots Threat to Microsoft," Dec. 3, 1998), you know the subject has widespread implications.

Yet, for all the attention, many commentators make a fundamental and flawed assumption that making source code available (hence the name "open source") is necessarily an anticapitalist move that could hurt a vendor's bottom line. "Somewhat akin to Coca-Cola Co. releasing its formula" is how The Post put it. Hardly.

Furthermore, The Post article perpetuates the myth that open source is a "concept once confined to the computing fringes, but one that has forged dramatic mainstream inroads in the recent months."

Unless one considers companies, such as IBM in the 1970s, to be part of the "computing fringes," the writer is wrong again.

I fear that today's movement that equates "open" with "free software" has got its semantics - if not its priorities - wrong. The current "economic doomsday for Microsoft" approach may result only in making Microsoft even more recalcitrant. A reality check is needed.

What open source advocates should do is lobby Microsoft to adopt the same stance vis-a-vis source code that IBM took with its mainframe software. IBM published the source code to its mainframe operating systems and virtually all subsystems, including the corporate jewel - the CICS transaction processing system.

As a systems programmer in the early 1980s, most of what I learned about the complex inner workings of these systems came from studying the microfiche of source listings that IBM made available to customers - for a small additional licensing fee. This information, along with heavily documented control blocks, message manuals and, in some cases, logic manuals allowed us to understand enough about our systems to keep them running.

Just because I could look at the source code - and attempt to learn from it - did not mean that IBM had given up its rights to charge a license fee for this intellectual property. IBM charged nontrivial fees to all users of its mainframe software. And charge it should - creating and maintaining that system software was an expensive proposition for IBM.

Did open source code ruin IBM's mainframe business? Absolutely not. To this day, customers willingly (if not happily) pay to use this software to run their businesses.

That policy was the catalyst for countless small companies that built everything from monitoring and diagnostic tools to entirely new subsystems, such as security and tape/disk management.

Most importantly, these entrepreneurial developers didn't have to secure IBM's permission to create their products. They didn't need to or want to steal IBM code, they simply needed to know how IBM's products worked so that they could interact with them.

One could argue that these "competitors" stole business from IBM, but I believe that they added far more than they took.

If Microsoft continues to keep source code secret, thus preventing others from doing things that Microsoft would never have the resource or the motivation to do, it will surely come back to haunt the company.

Tolly is president of The Tolly Group, a strategic consulting and independent testing firm in Manasquan, N.J. He can be reached at (732) 528-3300, ktolly@tolly.com or www.tolly.com.