Survey: Windows loses ground with developers

From a recent InfoWorld article:

Microsoft’s Windows platform is losing traction as a target for application developers in North America but still is the dominant platform, according to Evans Data survey results being released on Tuesday.
Free IT resource

A survey this spring of more than 400 developers and IT managers in North America found that the number of developers targeting Windows for their applications declined 12 percent from a year ago. Just 64.8 percent targeted the platform as opposed to 74 percent in 2006.

“We attribute [the decline] largely to the increase in developers beginning to target Linux and different Linux [distributions]. Both Novell and Red Hat are the two dominant ones right now,” said John Andrews, the CEO of Evans Data.

The arrival of Windows Vista likely only kept the numbers from being even worse. “I think Vista probably offset some of the decline,” Andrews said.

The share for Windows is expected to drop another 2 percent, to about 63 percent, in the next year, Andrews said.

The targeting of Linux by developers increased by 34 percent to 11.8 percent. It had been 8.8 a year ago, according to the survey. Linux targeting is expected to reach 16 percent over the next year.

Evans views the situation as a battle of Windows versus open source with open source maturing, Andrews said. Windows remains tops, though. “They’re still dominant, there’s no doubt about it,” said Andrews. Use of Windows on the development desktop remains steady.

The survey, featuring developers at enterprises and solution providers like system integrators, covered both client and server application development.

Evans Data said the shift away from Windows began about two years ago and is accelerating. Linux is benefiting as are nontraditional client devices. Evans Data also surveyed developer plans for such platforms as Unix and Mac OS but did not release those numbers.

Now, I dislike “surveys” for a variety of reasons, which I’ve detailed in the past. They’re easy to manipulate, so can often come out saying whatever it us you want them to. That being said, I think in the right context and given the right qualifiers, surveys can be extremely informative. Sure, the linked survey isn’t without issue, so shouldn’t be taken without a grain of salt (only North America for instance, doesn’t seem to take into account people using languages that don’t target a specific platform). But, it comes to a general conclusion that seems congruent with what most people are hearing. As an industry wide trend, developer support for Windows is declining. I think that’s a leading indicator and one I’d be worried of if I were Microsoft. It’s not just Linux that they’re losing to, though. It’s Linux and OS X and the web (a huge one) and platform independent technologies and mobile devices and, well you get the idea. They are fighting a rising tide. I think < 50% of developers targeting Windows will be the watershed moment. Given the current trends, that may only be a few years off.

A quick note on a related topic. I often hear people say that many Open Source applications have surprisingly large download numbers for Windows. From asking around and thinking about this, that’s not surprising at all. I’d say it should be expected, in fact. You see, the adoption of Open Source apps in many companies is a grass roots type adoption. Because of this, coupled with the fact that Windows is still by far the dominant corporate desktop, you’ll get huge numbers of developers downloading various Open Source apps for Windows so they can test them out on their workstations. Once they are comfortable with the product and have a proof of concept, they very often (but of course not always) deploy them on non-Windows infrastructures. Raw numbers alone don’t always tell the whole story. It’s about context.

–jeremy

Windows Vista – 6 Month Vulnerability Report

Jeff Jones, a Security Strategy Director in Microsoft’s Trustworthy Computing group, recently posted a 6 Month Vulnerability Report that compares Windows Vista, Windows XP, RHEL WS 4, Ubuntu 6.06 LTS, Novell SLED 10 and Apple OS X 10.4. Jeff has pointed out his potential bias, so I won’t even get into that angle. This report indeed does a better job than some from a methodology standpoint. For instance, he didn’t simply compare a default RHEL install, which includes a full Office suite and a whole host of apps not found in a default Windows install, with a default Windows install. He attempted to rip out the packages from the Linux installs that he perceived as being extra functionality when compared to a Windows install. This gives a much better baseline.

I’d like to simply offer a couple items that I think make reports like these a bit misleading. First, there is no standard definition of what a “Critical” or “High” security level is. It’s usually up to the vendor. It’s therefore possible that some vendors would rate nearly identical vulnerabilities with different severities. Second (and more importantly), we’re of course only looking at reported vulnerabilities here. Due to the Open Source nature of Linux, it’s much more likely that vulnerabilities will be discovered, reported and addressed. I’d contend that there are many more unreported vulnerabilities (which can be and in fact are still exploited) in proprietary software. If done again, another component I’d like to see added is average time to fix from time of first report. I’d be remiss if I didn’t point out that there are points that can easily be made for the flip side of the coin. The obvious one is that with a much larger install base, many more people will be targeting XP and Vista than other operating systems. In the end, statistics can almost always be made to say whatever you’d like.

–jeremy