Tidbits from the Eclipse Community Survey 2010
Join the DZone community and get the full member experience.Join For Free
the eclipse community survey is held every year by the eclipse foundation. you can find an accurate summery of the important trends in this post and the original report can be found here (pdf). after going over the report and the data, here are some notes i took.
pdt doubles cdt!
in the question "what is your primary ide?", jdt takes the lead by far with 53.7%, then pdt with 14.3% followed by cdt with 7.4%. this is a huge win for the pdt team. i think this trend will become even stronger. the new pdt 2.2 (to be released in the upcoming helios ) is a big improvement, especially in performance, and we will see more and more php developers upgrading to pdt. this is also an indication for the maturity of php.
interesting to see that only 9% answered that they use php as a primary language while 10.5% are using c/c++. almost 25% of the people answered that they allocate some time (over 0%) to php coding, compared with 31% on c/c++. the pdt package does include the java tools, so that might be a reason pdt is more popular.
scala anyone? ruby?
i've said it in the past: there are some vocal minorities. scala as a primary language got 0.4% and ruby got 0.3%. however, 4.4% answered that they allocate some time to scala, 4% to ruby. still, it is a minority.
the following graph shows the percentage of people who allocate any amount of time to a given language.
nobody does mobile?
there was a question regarding mobile application development. the results aside, 31 people answered it. that's about 1.8% of the people who took the survey. is there a strong bias here? about 13 people said they were doing j2me. they are probably the last :-)
eclipse version being used?
over 75% are using galileo. i wouldn't count on it, as it seems biased. people who answer the survey are people who read blogs, follow eclipse rock stars on twitter and frequent technology sites. these are innovators, hardly the mass market. it's suffice to see that over 7% are using helios milestone build, and that's before the rc builds (release candidates).
about 12.3% of the people in the survey are committers or belong to an organization which is an eclipse member. that's hardly balanced.
windows in decline?
from 2007 to 2010, windows as primary operating system went down from 74% to 58%. most users migrated to linux, some to os x. while i can see the trend myself, i think the 58% number is just too low. it is probably more subtle in reality. i know people from several companies that use eclipse, none of them (apart from google) use linux as their primary os. btw, if you were wondering, ubuntu is the leading the linux pack.
spring surpasses jee?
in the question "what is the primary server framework you use", spring got almost 20% and ejb behind with 19%. that is huge, but... there is no distinction between spring mvc and other spring components. so people who use ejbs with spring mvc (i'm guessing quite a few) are in the spring crowd. struts is still being used more than struts 2 and jsf: both never took off.
amusing anecdote: there are too many developers who don't understand the difference between a server side and client side.
redmine is popular!
i use redmine to track issues (i love it!) and it seems that i'm not alone. it wasn't even an option, yet, 47 people answered other and wrote "redmine". that's only 2.8% of the total answers, but, it's the 5th most popular after the jira, bugzilla, trac and mantis. it should be a first class citizen in the next survey and appear as an option on its' own.
people are happy, but...
there is an overall satisfaction. over 300 people chose to leave written feedback. i believe that every committer and every employee within the eclipse organization should take a minute and go through the feedback (the last question, 38, can be found in the raw data, links in ian's post ). there's plenty of criticism and a lot we can learn from.
personally, i found quite a few complaints around the issue of plugins. the game of mixing and matching plugins with incompatible interface and inconsistent quality seems to bother a lot of people. it makes me wonder whether there shouldn't be an approval process to plugins. you can publish a plugin, but it will only get the eclipse seal of approval once it goes through a certain process. i think that's something worth considering.
Opinions expressed by DZone contributors are their own.