On Feb 20, 2012, at 4:39 AM, MaXiujuan wrote: > Hi, Kevan, > > I am so grateful for your kindly and patient reply. These are all > valuable insights and comments, which mean a lot to me.
Happy to help. I should note that I work for IBM. > > > On Sun, 2012-02-19 at 16:28 -0500, Kevan Miller wrote: >> Oops. Forgot about your email... Apologies for the slow response. Hope >> this helps... > > Never mind. Please apologies for bothering you so much. And apologies > for so many questions below, please take your time and I am expecting > for any of your comments :P > >> I would characterize both periods as 'open source' periods. At all >> times, the Apache Geronimo PMC has been responsible for the management >> of the Geronimo project. Not Gluecode, not IBM, nor any other company. > > Thank you greatly for correcting me. Sorry, I took Geronimo in a wrong > way and now I realized that I hadn't learned it thoroughly when i heard > it together with "WAS CE"(i mean the community edition of application > server of IBM) on one of my school courses at the end of 2006. But i > still can't tell the difference between WAS CE and Geronimo, is WAS CE a > IBM version of Apache Geronimo? I am just curious about the relationship > of the two… Right. Apache Geronimo is a product of the Apache Software Foundation. WAS Community Edition is a product of IBM. WAS Community Edition is a redistribution of Apache Geronimo. There are some minor differences, between the two. But from a functional perspective they can be considered the same... > >> Gluecode was a company providing support for Geronimo. In 2005, IBM >> acquired Gluecode. In both periods there were commercial interests in >> the Geronimo project. There are commercial interests in most open >> source projects. > > I'm not sure whether this question is proper: What does the "commercial > interests" mean? Does it refer to the company's business model, for > example, sth. like IBM's strategy on open source application server, > e.g. in my imagination, IBM wanted to support Geronimo for its WAS CE or > Websphere? Ah, English with our overloaded terms. :) By "commercial interests" I'm using "commercial" with the following meaning: intended to make a profit. So, by "commercial interests" I mean that there are companies (or people) who are attempting to make money by their participation in / contributions to the open source project. Either via consulting/services, providing support, or other means (direct or indirect). > >> Organizations like the Apache Software Foundation >> take steps to insulate their projects from these commercial interests. > > What does "insulate their projects from these commercial interests" > refer to? I am very curious why to insulate, after all, it is quite > common in nowadays' open source projects with commercial sponsorship or > other kinds of involvement. Would those commercial interests lay burden > on OSS projects which would prefer to more free rapid technical > innovation? First, I'd refer you to some documentation about the ASF: http://www.apache.org/foundation/ http://www.apache.org/foundation/bylaws.html And perhaps most importantly: http://www.apache.org/foundation/how-it-works.html The ASF is a meritocracy. You can't "buy" your way into a project. You have to earn your way into the project, by contributing to the project -- creating documentation, answering user questions, creating bug fixes, creating new software features, etc. Once you've earned your merit, you'll become a project committer and eventually a PMC member. The PMC of a project manages the project. PMC members are expected to manage the project as "individuals", not as a representative of their employer. This doesn't mean that a company doesn't have influence on a project. They certainly do. However, the PMC is expected to insure that the community is operating in an open manner, to help resolve any disputes that may arise (e.g. I want to implement feature 'Foo', but you want to implement feature 'Bar' and we can't reach agreement -- the PMC is expected to help mediate our disagreement). If the PMC fails to perform our job, the ASF Board can (and will) disband the PMC (and ultimately could stop development on the project all together). Not all open source software projects are structured this way. Some are more open than others. IMO, the ASF is as open as they come. I haven't looked at the Contributor License Agreement (CLA) for JBoss. Nor have I looked at the by-laws for how their community policies. However, I expect that any contributions you make to jboss.org become the property of RedHat. If you are an independent contributor to a JBoss project as an independent contributor, I would not expect that my ability to influence the project would be equal to a RedHat employee. > > >> There was no new protocol for obtaining commit privileges. > > Thanks for the confirmation. > >> There was a period of time where the project was following a >> Review-Then-Commit (RTC) policy, rather than a Commit-Then-Review >> (CTR). This may have have resulted in slower development (and thus >> indirectly impacted new committers joining the project). > > Thanks, valuable point. > >> A number of factors may impact participation in an open source >> project. Release schedules, news articles, etc -- all can lead to >> increased participation by a community (e.g. I would expect more bug >> reports to be generated after a major release, and then taper off over >> the lifetime of the major release). > > Right, I did observe two increases of issue reports after Geronimo 1.0 > was released and after 2.0-M was released. Then the issue reports > tapered off, just as you said. Results also show the ratio of issue > fixed was quite nice, (averagely 80% of user reports were all got fixed, > and the remaining reports were invalid or duplicated..), suggesting > developers in Geronimo project have done a very good job:) :) Thanks! > > >> The RedHat acquisition of JBoss is quite different, IMO. Personally, I >> would not equate an ASF project with an open source project like >> JBoss, which is essentially controlled by a company (JBoss and then >> RedHat). > > Right, Thank you, now i know more about the difference between JBoss > project and Geronimo project. > >> The departure of employees following an acquisition is not unusual. >> JBoss was "commercial". RedHat is "commercial". RedHat is also >> "larger". I expect there were a number of reasons why JBoss employees >> might have left: cultural, philosophical, and economic. > > Exactly, it should be a problem when combining open source development > into a commercial environment, not existing in Geronimo, isn't it? Well, JBoss was a company. They developed JBoss and were attempting to make a profit. It was a "commercial environment" prior to the acquisition. I'm sure it was a different environment (e.g. smaller, more entrepreneurial, more technical freedom, etc). However, both companies were attempting to make money via the creation of JBoss. > >> I don't know what you mean by developers "leaving" (or how you're >> measuring this). I wouldn't want to discuss any individuals by name >> (or otherwise) on a public mailing list. 2003-2005 is, of course, a >> shorter time period than 2005-2010. People become involved in open >> source projects for a number of reasons (and also stop participating >> for a number of reasons). A company's interest in sponsoring employees >> to participate in an open source project could certainly play a role >> in this. > > Currently I just focus on the commit activities of developers. By > "leaving", I mean a developer stopped committing code then she will not > included in my investigation. That is, I calculated the duration between > their fist SVN commit and the last SVN commit. There are still some > details to deal with, for example, if a developer finished the last SVN > commit over 1 year by the date I completed the data collection, I > assumed the developer had "left" the project. OK. Thanks for the information. > >> Companies may ask their employees to participate in open source >> projects. IBM has (and continues to) sponsor participation in the >> Geronimo project. Some companies may have sponsored participation in >> Geronimo, but later decided to allocate their resources to other >> projects. Same is true for individuals. >> > Would you mind if i ask some questions about the "sponsor > participation"? Does that mean companies should give some money to ASF > for Geronimo? Or they just needs to allocate their employees to > contribute to Geronimo? The latter. Companies do donate money to the ASF (see http://www.apache.org/foundation/thanks.html). These contributions are used by the ASF to run the foundation, maintain infrastructure (hardware, software, networking, power), legal, etc. However, this money is not used to directly fund software development on ASF projects. Companies are able to ask their employees to participate in/ contribute to a particular project. > > Another question is: Do the employees from IBM full-timely work on > Geronimo? It seems those developers were quite productive measured > through their SVN commits. Yes. > >> There are no process changes regarding Jira. You may find a >> correlation of increased Jira reports with major Geronimo releases (a >> period where more people may be using Geronimo, also a period where >> there may be more bugs in the software). > > Right, refer to the above observation that there are two big increases > in the numbers of user-reporters and their reports right after releasing > Geronimo 1.0 and 2.0. I think these two increases show both of the cases > you described: both more people using new release of Geronimo and there > were many issues in the new release, right? Correct. > >> Use of Jira is encouraged (though not strictly required). It provides >> a way of tracking new features and functions that are added to >> Geronimo. It's also not unusual for developers to creat a Jira issue >> based on a user's report of a problem in an email, IRC message, etc. > > Thanks for the details. > >>> >>> I believe what Geronimo is doing has great implications for software >>> engineering, e.g, how to build a better community, what are the best >>> practices to attract participators. I would greatly appreciate any >>> comments you might have. >> >> Hope the above helps. > > I appreciate your help very much. They are quite precious for me. > Sorry to bother you again. No problem. Good luck with your research! --kevan
