« Following "the Demise of Traditional Software" | Main | No Red Meat at OSCON... Yet »

If you’re into numbers, take a look at a couple of interesting and not necessarily conflicting statistics about open source software (OSS) put into circulation in July: Linux up, Apache down. Among other very detailed findings, 10,000 Alfresco community members said during calendar quarter two that they evaluated OSS using Windows and then deployed on Linux if they liked what they saw. Based on another survey, Glyn Moody of opendotdotdot is opining in Linux Journal about why Apache is losing a lot of web server software share to Microsoft Internet Information Server.

According to a press release, the Alfresco Open Source Barometer Survey showed that Windows is increasingly “a popular evaluation platform” for OSS but that “most enterprises use Linux when they go into production.” That’s pretty consistent with decades-long Windows-UNIX market activity. I haven’t looked at the survey details (they are available at Alfresco’s site) but some bloggers are reporting that evaluation equals development. If so, that’s a different trend; typically more developers deploy to the stack they developed on than not. If that’s not happening with Windows-based OSS, the trend bears watching.

The Alfresco survey also asks users about their preferences in operating systems, application servers, databases, browsers, and portals but remember that the results are not a good barometer of the IT industry, just of Alfresco community members. It doesn’t matter what the “n” is in surveys. You could Alf Landon why that is but I think this is the first U.S. presidential cycle since 1932 in which he is not running for president.

As for the other survey about web server software, Glyn’s source is Netcraft. I use the very imprecise terminology “a lot of” in describing the numbers (supposedly 15 points of web server software share lost by Apache to Microsoft Internet Information Server in the last year; 2 million servers added by Microsoft in the last month) because I can’t gauge statistical significance despite over 125 million “n’s.” I quickly looked at Netcraft’s site for their methodology but couldn’t find it. I believe it is based on some kind of webcrawling/pinging thing but why it picks up some servers and not others could skew the results.

So take the “Jerry McGuire” advice we give clients at Research 2.0 (see link to right): If you are interested in what your peers are doing, follow the money (revenue), and/or count the paying customers coming through the turnstyles.

Posted by dennisb in OSS Business Issue |Digg This|Add to del.icio.us

Trackback Pings

TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/2154