A software census would answer open source software (OSS) questions - Open Source Software Up the Stack

« Elemenope, a great OSS project name and a good way to use OSS integration servers for BPM, supply chain automation | Main | At the OSI "license-discuss list," you can hear from the open source software (OSS) pros »

January 22, 2008
A software census would answer open source software (OSS) questions

OpenLogic on January 22 released a list of its customers' most popular open source software (OSS) packages. Two interesting findings were that the Apache license is more common than the GNU General Public License (GPL), and that Hibernate and Struts topped the list of most popular packages "with more than 71 % of customers using each."

I have a call into OpenLogic to see if I can get some information on the survey's methodology. It does not surprise me that Apache is more popular than GPL. Hibernate is distributed with a Lesser GPL license (LGPL) but many of the rest of the leading projects on OpenLogic's lists are licensed under Apache (I think I can say most but I don't feel like doing the detailed research).

But why are the Apache HTTP server and Tomcat packages so relatively low on the OpenLogic list? I suspect it is something unique to the OpenLogic customer base. Similarly, the hit of last week's gala, MySQL, made the list but I do not see many other products that I would expect such as Alfresco, Firefox, Mule, and SugarCRM. I'll update this if I find out anything other than that the sample was skewed by the demographics of OpenLogic's customers.

But you can't blame OpenLogic. This press release reminds me that beginning next quarter we can hopefully expect more broad-based census information on OSS if users particpate. As discussed in this podcast with Kim Weins, Marketing VP at OpenLogic, the company is hoping to get wide participation in a cenus among anyone--OpenLogic client or not--who will use its OSS Discovery software, which can be downloaded at its site.

So for example, we will be able to find out with more statistical accuracy the answers to questions such as:
-- How prevalent is the LAMP stack vs. the WAMP stack (and WAOP and LAOP stacks for that matter?)?
-- How much OSS is truly free and how much is just open?
-- How quickly are open source ESB's permeating enterprises?
-- And many similar questions that users, managers, developers and investors care about and that I have written about in the Features section that accompanies this blog. In the long run, what would be most useful is a census that didn't make the distinction between open and closed source. That would settle a lot of arguments.