This year I‘ve talked a lot about opening up the JCP. I first did a round table at QCon with Rod Johnson and Patrick Curran (chair of the JCP) back in February. Then at Java One (with the same folks) and in May Patrick Curran came to the Paris JUG … and now a BOF at Devoxx with Corina Ulescu about this sensible topic.
The questions are always the same : should the JCP be more open, and how ? To the first question, I think that most of the people agree. Yes, the JCP should be a more transparent organization. But to which extent should we open it ? And if we find a solution, do we have to formalize it and impose it (that means changing the JCP process). Basically, a spec lead can decide of the tools he/she wants to use. If he/she wants a public mailing list, that‘s fine, if not, that‘s fine too. The JCP does not impose a spec lead a way of communicating. So, at the moment, there‘s a bit of everything : closed JSRs (no wiki, no public mailing list…), opened ones (open source reference implementation, open mailing list…), and anything in between. But which model to choose ?
I have to say, I‘m between two waters here. I‘m expert member on JSR-316 (Java EE 6), JSR-317 (JPA 2.0) and JSR-318 (EJB 3.1), that‘s a lot of work. Every day I receive emails (and not trivial ones) that I have to read, sometimes compare with old revisions of the spec, answer back, ask questions, look for information on the web and make up my mind on hot decisions. That means, having a closed mailing list is already enough work. If we leave it completely open, the traffic would be even greater and, as Brian Goetz pointed out during the BOF, that would also mean more noise (people sending emails on topics that have already been taken care of or that they don‘t understand). As I‘m an individual (I don‘t have a company to pay me for reading and writing emails on a specification), I would not be able to follow that pace.
During the BOF, Brian Goetz pointed out that in the word “expert group“ there‘s “expert”, and we can‘t spend too much time on explaining in details trivial things on a public mailing list. Writing a spec is already a lot of work. Kirk Pepperdine, on the other hand, wondered why we couldn‘t have the entire public mailing list being the expert group.
For me, I prefer the idea of closed mailing list. But there‘s one thing I‘m really frustrated about the JCP, is not being able to easily observe a JSR. To observe a JSR, to have to send an email to the spec lead (who‘s already under lots of work) and he/she can decide then to add you or not. I think anybody should have a read-only access to any JSR private mailing list. Like any other mailing list, you should be able to subscribe to any JSR. That would bring many benefits:
- the expert member would know that people are reading behind their shoulders and avoid political emails, sometimes it happens
- people would have more knowledge on topics and would write less non-sense articles
- if you want to answer back to the list, there‘s always the firstname.lastname@example.org that you could use
- if you strongly disagree and want to make your point, blog about it, write about it (with the knowledge that you get from being an observer)
So my two recommendations are :
- let anybody easily become an observer
- if one person follows the list and gets interested (by writing and bloging), let him/her be part of the expert member
So, for me, open up the JCP, but not completely (I‘m not in favour of having public mailing list). What about you ?