Eclipse gains market share in 2005

Each year BZ Research (publishers of SD Times and Eclipse Review) do a ‘Java Use and Awareness‘ survey of the SD Times readers. One of the question is ‘What java development environment do you use?’ They have been doing this since 2002, so you can see some interesting trends.

Not sure how I missed their most recent survey but it is from December 2005. They have given me permission to blog the results of the Java IDE question. (btw, it is also referenced in the Feb. 1, page 12, edition of SD Times.) The good news is Eclipse grew 9% in 2005 and now stands at 65.1%. If you add in the IDE’s that are based on Eclipse, the number gets even better. Interesting, one IDE that is claiming great momentum seems to have no growth.

Measuring the success of Eclipse or any open source project is always a challenge. Therefore, market research like this always raises my interest. I also like the BZ Research survey for a couple of reasons:

  1. I guess the obvious reason being Eclipse is on top 🙂
  2. The sample size is large, in this case 621, and the sample is derived from the readership of SD Times; which I believe provides a good sample of the IT industry. Do people agree with this assertion?
  3. It provides trend data going back to 2002, so you can see what changes are occurring.

btw, sorry for the poor graphic quality.

13 thoughts on “Eclipse gains market share in 2005

  1. Leif,

    First, I want to apologize to Roumen. The intention of this blog post was to focus on the postive growth in the Eclipse community, not a slight against NetBeans. Roumen has always taken a positive voice in his blog and it is always my intention to do do here.

    I was not aware of the numbers Roumen has posted and I am glad he has made them public. What I do find perplexing is why the BZ Media survey shows no growth for NetBeans. I do believe the BZ Media research is a statistically significant representation of the SD Times readership but I don’t understand the difference.

  2. Missing growth? Studio Enterprise and Creator (which are based on NetBeans) both went from n/a to about 18% [combined].

    So… The awareness of NetBeans based IDE’s doubled in one year.

    That isn’t terrific, but it certainly isn’t “no growth”.

    I think both of these products “hit” the update center… So, they may account for some of the stats that Roman presented, too.

  3. Ian, apology accepted. I will keep positive voice in my blog, unless I’ll have to correct some numbers like these 🙂

    Vince, the numbers from my blog contain only NetBeans Update Center visit data. The other Sun’s IDEs are measured elsewhere. Their usage is also growing – no surprise since they are much newer than NetBeans.

  4. Only a blind men could say that NetBeans has not grown, the only answer is if you go ask the users of eclipsezone how many have downloaded and used NetBeans in the last year the answer will not be zero or -0.2%.

    NetBeans has grown exponentially this year, how do i know i am an active member of that community and i saw it grow in front of my eyes. The community grow stronger and larger, the plug-ins keep on coming and the tutorials, flash movies, etc… all that activity makes me believe that NetBeans is the great winner of 2005/2006, even thou awards say the contrary… some people like to work in the dark 😉

    NetBeans has made a leap this past year, while other IDEs have made small steps, and are playing catch up in many areas (Matisse anyone?).

    Is this the same SD Times that has a special feature article with the title “Spotlight on Eclipse”? Hum interesting…

  5. Vince,

    I think Roumen answered you guestion about including the Studio numbers in with NeBeans. Also I would caution anyone from adding numbers across multiple IDEs, since a respondant can select multiple IDEs. Therefore, I would think NetBean users would be more likely be using Sun Studio and Creater, just like IBM Websphere users would use Eclipse. This is why I did not add up the products that use Eclipse as their platform, to get to a bigger Eclispe number. 🙂

  6. What I do find perplexing is why the BZ Media survey shows no growth for NetBeans. I do believe the BZ Media research is a statistically significant representation of the SD Times readership but I don’t understand the difference.

    No mystery really. The data I’ve seen suggests SD Times readership (not to mention “Eclipse Review”) are essentially Eclipse users and their coverage caters for Eclipse users. So while their results may be useful for the Eclipse community, the survey pool has a high degree of pre-selection that perhaps makes the very inclusion of NetBeans in their survey something of an error.

  7. “I believe [this survey] provides a good sample of the IT industry. Do people agree with this assertion?”

    No. Choosing a random sample is critical to the statistical validity of such a survey. I seen no evidence that this is a random sample. In fact, there is much to suggest otherwise.

    What this means is that we probably have no more knowledge about the relative sizes of the user bases for Java IDEs after this survey than we did before it. In fact, the stats might be so bad that even the ranking of the popularities could be wrong.

  8. Everyone is missing the point. It’s about growth. Even if you want to play “conspiracy theory” 101, then are you suggesting that BZ Media is also faking growth? Come on. If you want to claim the %’s are skewed, then I might buy it. IF you want to claim growth is skewed? Nope, sorry.

  9. Simon,

    I disagree with the statement that SD Times caters to the Eclipse users. SD Times has significant coverage on MS Visual Studio, Java, embedded technology, etc. There coverage caters to a wide audience. I think it is more accurate to say it doesn’t cover NetBeans.

    I find a lot of people are trying to paint SD Times as an Eclipse specific magazine, which I don’t buy. I do agree that they don’t cover NetBeans.

  10. Ian,

    I don’t disagree that SD Times covers different aspects of the industry. The point is that in order for this kind of survey to have any meaning, the sample of developers needs to be *random*. And it just doesn’t look to me like it was.

    Of course, people will always believe what they want to believe. So, Eclipse supporters will believe that, despite any errors in the design, this survey shows Eclipse is gaining market share. Whereas, Netbeans supports will believe that the survey doesn’t accurately reflect Netbeans usage.

    The truth is: if the sample wasn’t random, the survey doesn’t show *anything* to *anyone*. Not absolute sizes, not growth, not rank order. Nothing.

    No conspiracy theories required. Just a typical magazine survey.

  11. Simon,

    First let me correct my previous comment, it was intended for webmink, who is actually Simon Phipps from Sun, so sorry for the confusion.

    I am not stats expert so I would have to trust you on what you say. Do you know of any survey that would fit your criteria.

Comments are closed.