 On surveys and statistics
One of the key pieces of news in the Eclipse world recently was the staggering survey results from the Eclipse foundation. You can see the press coverage here, and you can read some of the less formal insider information here. Last but not least, Riyad (my
partner in crime) covered the survey here at EclipseZone, with helpful links to exclusive EZ content.
This is the second year in a row for this survey, and the numbers as reported are certainly impressive. Across the board, various Eclipse features have gained adoption, and Eclipse itself is only getting stronger (if that is even possible). As an example, in 2005 when asked the question "If Eclipse IDE or Eclipse-based products are being used
in your company, about how many developers use them?", 4% of respondants chose 'More than 1000', while in 2006, 24.9% chose 'More than 1000'.
On the positive side of things, I certainly recognize the value in surveys like this. Eclipse planning and requirement gathering needs to look at how Eclipse is being used, and perhaps more importantly how Eclipse is not being used to help guide their compass. At the same time, it's valuable to look at the numbers in retrospect to get some idea of what approaches, marketing, and planning worked, and also what of the same didn't work.
The waters get a lot murkier, however, when you start using these surveys as a marketing tactic and as some end-all be-all of Eclipse's position in the marketplace as a whole. After all, this is an Eclipse 'funded' study, it doesn't consider other competing products, and it's not a long-term representative sample. It is probably premature to jump to conclusions by applying these numbers to other data sets in the world, making correlations where there are none.
Also, let us not forget that Evan's Data, the company that Eclipse is sponsoring to conduct this survey, is the same company that created such a fury only two months ago by running a survey of popular IDEs that ignored some of the most popular for Java programmers, such as Intellij IDEA.
Now, don't get me wrong - I'm extremely pleased that, when given an apparently representative sample, Eclipse usage has climbed on almost all fronts - it's a testament to
the work being done by the foundation, and it speaks good things for the strength of the platform and the tools as time goes on (a win-win for Java developers as I like to call it). I'm also not saying that the Eclipse foundation is guilty of mis-using these numbers; in fact they have been very controlled with this information, and haven't (as far as I've seen) jumped to any conclusions. I just cringe at the jump-to-conclusion office-talk that numbers like this are bound to create. After all, remember what they say: There are Lies, damn lies, and then there are statistics.
Anyway - see you at the water cooler.
Until Next Time,
RJ Lorimer
rj@javalobby.org
|
Point & Click Clustering - Free Download Available Now
EZ News Corner
