Discussions

News: JBoss Usage Survey Closed, Results Available

Bookmark and Share Get thread feed

  1. JBoss Usage Survey Closed, Results Available (14 messages)

    The JBoss usage survey announced in collaboration with Richard Monson-Hafael has been closed, and the results are available from SurveyMonkey, which provides some filtering capabilities too. Three hundred sixty-four people took the survey; the results indicate that JBoss has the lead among the survey respondents, and Hibernate and the App Server itself are the important bits. It's not clear whether this is an unbiased poll; obviously, it was on JBoss usage, so non-JBoss users might have felt less interest in participating. There were no real surprises among the results, although the percentages were interesting. Most respondents indicated they were end users; at small companies; and were end users. However, it's unclear whether these percentages only look like correlations, or whether they actually were correlations. (It's true! Just because 70% say "A," and 70% say "B," there's no way to validate from just those results whether the same 70% said "A" and "B.")
    Reply to this

    Threaded Messages (14)

  2. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    However, it's unclear whether these percentages only look like correlations, or whether they actually were correlations. (It's true! Just because 70% say "A," and 70% say "B," there's no way to validate from just those results whether the same 70% said "A" and "B.")
    It's sad that you have to explain that but you probably do. The raw results are there if you click on the view detail button in the upper right hand corner of the window. You could determine whether these were correlated that way. There weren't many responses. I doubt this survey was statistically significant.
    Reply to this Reply to original

  3. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    It's not clear whether this is an unbiased poll; obviously, it was on JBoss usage, so non-JBoss users might have felt less interest in participating.
    There's really no way to make such a poll unbiased - so you just have to make it clear what the sample is based on "TSS visitors between a certain date that chose to respond to a survey on their JBoss usage". There was no point in non-JBoss users participating in a survey about their JBoss usage. But that's fine. It's clear that it's a survey for and about JBoss users. And the results were interesting, even if the sample size was very small. The bias at the moment toward smaller end user companies is normal for open source at the company's stage of development. I'd guess that RedHat is hoping the aquisition will help JBoss break into the upper end of the corporate world - I'm sure it will. PJ Murray, CodeFutures Software Data Access Objects and Service Data Objects
    Reply to this Reply to original

  4. You can make correlations![ Go to top ]

    However, it's unclear whether these percentages only look like correlations, or whether they actually were correlations. (It's true! Just because 70% say "A," and 70% say "B," there's no way to validate from just those results whether the same 70% said "A" and "B.")
    That's not true. There is a filter button that allows you to, for example, see all though users you chose A and B or A and not B. Etc. The user population is intended to be JBoss users, we have a question in there that allows us to filter out non-JBoss users. The sample size is smaller than we hoped, but its only one data point in a larger research project so while its not perfect it is helpful in understanding how JBoss is used by the respondants of the survey. Thanks everyone! Richard
    Reply to this Reply to original

  5. Re: You can make correlations![ Go to top ]

    It seems that if the intended population is supposed to be JBoss users that there are quite a few questions about the competing platforms. Sounds biased to me.
    Reply to this Reply to original

  6. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Hibernate and the App Server itself are the important bits
    Based on the way I responded to the survey questions that were worded, I would tend to doubt the validity of the above statement. I responded that we use Hibernate. But we have been moving our efforts away from Hibernate and to iBatis. So the answer that we use Hibernate is there, but it is without conviction. As for the usage of the JBoss Application Server. That choice is not one of acutal need, but one of "It's there and I can use it with impunity". Most of our applications really require Tomcat, JNDI datasources and occasional Stateless Session EJBs and JMS/MDBs. The real important question is only useful when combined with two other ones. "What are the most important Java EE products used by your organization in production?" combined with "Which best describes your company" and "Which best characterizes your company" would give the best view of JBoss in the overall market. Almost every organization I know has at least one application running on JBoss. Why? Because of the licensing. Smaller firms use it because other servers may be cost prohibitive. Larger organizations have used it because the applications to be deployed on the server would not have been able to get approval because they weren't able to provide a positive ROI if deployed to other environments. Therefore a large organization may use JBoss in production. But their mission critical and highly scaled applications are running on other JEE App Servers. I don't know if my assumptions are true. But the confimation (or refutiation) of this theory is what interests me. I want to know if JBoss is part of some serious mission critial work. Or is it residing in the fringes of the overall enterprise culture. John Murray Sobetech
    Reply to this Reply to original

  7. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Larger organizations have used it because the applications to be deployed on the server would not have been able to get approval because they weren't able to provide a positive ROI if deployed to other environments.

    Therefore a large organization may use JBoss in production. But their mission critical and highly scaled applications are running on other JEE App Servers.

    I don't know if my assumptions are true. But the confimation (or refutiation) of this theory is what interests me. I want to know if JBoss is part of some serious mission critial work. Or is it residing in the fringes of the overall enterprise culture.
    As long as you don't use persistent messaging. If you need to use persistent, exactly-once, XA messaging, you need to replace JBossMQ (but with what ..?) and deploy JBossTransactions, and hire people to support it. I repeat: the so-called "fast, in-memory" Jboss transaction manager does not have a recovery log. That answers the question about "mission critical". This will all be moot soon since Geronimo has a recovery log and ActiveMQ, and will eventually get clustering and JEE5 support, and that will be the end of it, just like Tomcat ended the question of what servlet runner to use. Guglielmo Enjoy the Fastest Known Reliable Multicast Protocol with Total Ordering .. or the World's First Pure-Java Terminal Driver
    Reply to this Reply to original

  8. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    I repeat: the so-called "fast, in-memory" Jboss transaction manager does not have a recovery log.
    Installation of the JTA component of JBossTS is fast and simple and fixes this.
    Reply to this Reply to original

  9. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    This will all be moot soon since Geronimo has a recovery log and ActiveMQ, and will eventually get clustering and JEE5 support, and that will be the end of it, just like Tomcat ended the question of what servlet runner to use.
    So are you saying that JBoss will be replaced by Geronimo or that they will both be commoditized? Do you feel that JBoss is in the mission critial JEE server space that BEA/IBM/Oracle troika is? Or is JBoss a fringe player? Do you anticipate that Geronimo will be the Apache/Tomcat of the JEE space. Where Apache has replace all other Web Servers and Tomcat has replaced all other Servlet Containers?
    Reply to this Reply to original

  10. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    I didn't like quetion 4 because it did not offer Tomcat as a choice other than "other." Question 2 does give it as an option though. Does this mean that a JEE app server has to have an EJB container? It would have been nice to see what backing database was used by respondents. I take away that the JBoss user base is small and fragmented. Wonder how this correlates to Redhat demographics?
    Reply to this Reply to original

  11. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Andrew Clifford :
    Does this mean that a JEE app server has to have an EJB container?
    By definition, yes. geir
    Reply to this Reply to original

  12. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    What Jboss App version are they using?
    Reply to this Reply to original

  13. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Very interesting discussion I have only one comment made up of one word......WebSphere
    Reply to this Reply to original

  14. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Very interesting discussion I have only one comment made up of one word......WebSphere
    Reply to this Reply to original

  15. Re: JBoss Usage Survey Closed, Results Available[ Go to top ]

    Very interesting discussion I have only one comment made up of one word......WebSphere
    Have you used it at all, or you are just fascinated by the big name of IBM? I, personally, have never seen a person who, given a choice, liked WebSphere. It's arguably, absolute worse app. server.
    Reply to this Reply to original