Numbers may not lie, but survey respondents sometimes do. That came to mind when I read that Linux Gaining Strength In Downturn. IT managers can answer a survey any way they want, but the real test is whether they commit money to Linux, or divert money from Microsoft to Linux.It's not that people intentionally answer these surveys deceptively; they often just don't know what they are planning to do until they get into a meeting, analyze the facts, and need to commit to a decision. In 1989 through 1991, the IT magazine I worked for at the time would do regular surveys that showed many IT managers planned an organizational migration to IBM's OS/2 within 18 to 24 months. That 18-to-24-month answer stayed about the same over those years, indicating that IT managers were in reality deferring the decision. In retrospect, we all know OS/2 migration day never came for most organizations.

I think the "OS/2 soon" phenomenon is behind the Linux-friendly answers to this survey. Let's start with the desktop numbers: "Sixty-three percent said they will increase their use of Linux on the desktop by more than 10% this year, although such an increase would still probably represent a miniscule share of all desktops." Yes it would. So perhaps this number is realistic strictly because a 10% increase of a small base is certainly attainable, but not (yet) significant.

The projections on the server side seem wild, though: "Forty-nine percent said they expect Linux will be their primary server platform within five years." The Gartner numbers I could find showed that Windows had 67% of the server OS market in 2007, with Linux at 23%. What we don't know is how many of those 2007 servers shipped into "hybrid" companies that use both environments, and which went to companes that use one OS exclusively. For that five-year plan to come true, though, you'd need to see quite a few Microsoft-centric companies abandon Microsoft. It's not going to happen.

With the economy taking a serious downturn, IT managers are being called upon to find ways to reduce costs. Linux is free and Windows licenses aren't, so at that level it seems like a money-saving measure to make the Linux move. But there are so many other issues that drive up the cost of switching. A Microsoft-centric shop doesn't just have Microsoft Windows, it likely has Microsoft SQL Server, SharePoint, Office, Internet Information Server, and so on. The staff is trained on how to maintain and use those applications, and internal company applications depend on them.

No matter how much the survey questions indicate that IT managers want to move to Linux, reality and switching costs will prevent many of them from doing so. Very few IT managers will be able to make the case that it's worth increasing their budget to cover the one-time costs in the middle of a recession, with the supposed payback coming a few years down the road.