On May 31 in a blog entitled In Open Source and Cloud Computing, Beware of the Curve Ball, I suggested you "check twice before you cut" (the standard advice for carpenters) when it comes to what you read about enterprise software in the mainstream media (MSM). It's nice that publications such as The Economist and BusinessWeek are paying attention to your profession but they are treating the facts so superficially (because their primary audience is not IT managers and staffs) that their conclusions are misleading.to your management.
The real place you need to check the facts is in blogs. The headline in a June 4 blog post by Red Hat's Michael Tiemann entitled "U.S. CIO Vivek Kundra Advocates Open Source Software" is a great example. Even though that headline was repeated in numerous places on the Web (and Mr. Tiemann had based his headline on an article in Federal Computer Week), it appears that Mr. Kundra had done no such thing. The saga of the disappearing and weak-kneed Obama-administration open source endorsement on FCW is explained on my IT Investment Research Web site.
The thing that really bothered me in Mr. Tiemann's post was an array of charges that journalists are unfair in reporting on the open source software culture and its defining terms and conditions (Ts&Cs). Implicitly in his post (because of the use of the term "FUD") - and explicitly in many similar blog posts - these charges are accompanied by conspiratorial allegations that journalists are in the bag for the so-called proprietary enterprise software vendors. I blogged about that absurdity as it relates to analysts on August 26, 2008 and the same applies to the technology journalists I know.
A key Tiemann line was his last sentence:
"I propose more talk about the specifics of what people are actually doing and less time talking with those who, contrary to all evidence, claim it cannot be done."
Assuming that by people he means you in the IT managerial and staff class, I think that's a great idea. The facts put the lie to his charges against journalists.
Based on publicly available hard evidence, as opposed to blogoblather, here's what you are doing from an IT-market revenue perspective:
- You are spending around $3 trillion a year on IT products and services worldwide to support your enterprises, organizations and governments.
- You are spending less than 10 percent of that - less than $300 billion worldwide annually - on software-specific product and servicces.
- Around 90 percent of that $300 billion involves adding users to, upgrading, or maintaining existing enterprise software (as opposed to buying new-new software).
- You are spending less than 15 percent of the $270 billion a year (90 percent of the $300 B) or about 1percent of the $3 trillion - less than $40 billion worldwide - on software-specific services tied to software released with restrictive open source software Ts&Cs (software with non-restrictive open source Ts&Cs is indistinguishable from any other software so it is hard to measure but most likely is also a similarly small number).
So from a revenue perspective, the open source software movement is a freckle on the back of a gnat on the back of an elephant. That's why journalists don't cover it the way the open source blogblatherers would like. It has nothing to do with the so-called proprietary software providers (which of course is also a favorite open source blogoblatherer concept with no basis in reality).
But does this mean the open source movement is a failure? Or does it mean that the vast majority of you in IT are stupid to have not followed the advice of "open source" companies such as Mr. Tiemann's Red Hat (RHAT) to choose only restrictive open source Ts&Cs? Or does it mean there should be another metric other than revenue applied to open source software?
The answer is "none of the above." Item 4 in my above list, although a popular way of looking at such things, is misleading. There is no open source software market so you can't effectively measure it. Because of the legal Ts&Cs aspects of the open source culture, I can statistically calculate an "open source market" (and I have if a client asked me). But I also tell the client that such a measurement is not an effective means of market research because you in IT tell us that very few of you actually acquire software or any technology because of the way it is licensed.
You acquire enterprise software because of what it does. You run MySQL and Apache on top of Windows. You run SAP and the Oracle RDBMS on top of Linux.