James Governor's Monkchips

Open Source Java Will Mean Developer Churn:Zero Sum Technocracy

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

SunMink : SunMink

Simon Phipps:

“One of the biggest questions on the table right now is, how can we open the code, invite participation from everyone with the interest and the skills to do so, and still respect the work and ongoing contribution of the existing developers? Many of them have spent a decade or more making the Java source code excellent. We can’t just dump the code on the street and invite applications for the role of owner for each module. That would be hugely disrespectful, dishonouring the developers who have devoted themselves to the code. And yet neither can we just assert that no help is needed.”

Simon gets to the heart of a matter here, and the fact is, Sun is going to have to break some eggheads’ shells. “Innovation happens elsewhere” – well, now more fresh ideas can be driven into the core platform, which will mean upsetting some people. I think upsetting a few Sun engineers however should do Java no harm at all.

This is an area where the “developer self-interest in writing code they want to use” is tested. When people don’t want the code you wrote, after ten years of stewarding it, what happens then? A corrolary of freedom to leave…

Tags: , , , ,

9 comments

  1. And hardly a problem unique to Java’s development team. I know it’s not quite as ubiquitous but I’m sure some Ingres RDBMS developers felt the same way…?

    You say Simon gets to the heart of the matter: I say, “well, durrr…”

    Opening the process is only disrespectful if you botch the process, or the people concerned have, at the back of their minds, the idea that opening the process means that they have failed, or worse still, that open development processes are in some way inferior to closed ones.

  2. is durrr the same as duh?

    I do believe the human elements are underdone in many commentators discussions on the subjects. the bbc would like to open source its content as well, but the question is, whose IP is it anyway?

  3. “Many of them have spent a decade or more making the Java source code excellent. We can’t just dump the code on the street…”

    Sounds like Mr. Phipps isn’t quite a true believer in Open Source yet!

    For me it is a no-brainer: Java is a perfect target for open sourcing – it is mature, well defined and widely used. I simply can’t see how Sun can commercially justify bearing the cost when they can gain kudos and ‘free’ development effort by going open source.

  4. yes make a fine point ronan. one thing that does interest me is that Simon doesnt talk to people *outside* Sun owning any of the code. That surprised me a bit.

  5. Ronan: What process are you suggesting for migrating the commit access for all the modules of Java SE?

    James: There are no current non-Sun employees who are the maintainers of any of the modules of Java SE, what is your point?

  6. which point Simon? what is my point about churn? or what is my point that you dont talk to non Sun contributions- i assume you mean the latter. its just surprising that a piece of code contains only contributions from employees- that there is no “orphan” code, or even licensed code in there. many code bases at commercial software companies include all sorts of rights issues. So my point is i guess that if Sun only has internal rights issues to work out, then it shouldnt be as hard as open sourcing solaris, for example.

  7. Simon,

    Rereading your original posting and then my comment, I can see that my first remark probably came across as a little glib and possibly ill-judged. There are clearly huge challenges in moving a large code base into the open source world – as somebody who has come from a software development background, I appreciate how hard it may prove to be.

    However, I am surprised about your focus on the existing developer community inside Sun being disrespected by open sourcing their work. Two dimensions to this:
    – Firstly, while I can appreciate that it can take some time to get comfortable with the concept of open sourcing – it can hardly regarded as a negative report on the success of Java or the success of the team responsible for that success!
    – Secondly, I would imagine that the same developers would naturally fill the leadership roles in a future open source version of Java. And in today’s world, I suspect that they would get greater kudos and respect from filling those roles than from their roles inside Sun which may be somewhat invisible to the greater development community.

    All of which made me wonder whether there is a concern with some in Sun around the issue of open sourcing in general rather than open sourcing Java (and I clearly don’t know your views on Open Sourcing and would not wish to comment on them). This would be clearly surprising given the very visible commitment from Sun to Open Sourcing – hence my first comments.

    Ronan

  8. There is at least one counterexample which by all accounts has been deemed a success, namely the java.util.concurrent code maintained by Doug Lea. (http://g.oswego.edu/dl/classes/EDU/oswego/cs/dl/util/concurrent/intro.html).

    This is the same code used by Sun as I understand it (and we’re using it in Apache Harmony as well…), and this model of a non-Sun domain expert maintaining the code has worked out well so far. And as a bonus, there’s a symmetrical relationship with the community wrt copyright.

    (And of course, the door is always open at Apache Harmony for Sun’s implementation of SE if they need a transparent community with a level playing field… 🙂

    geir

  9. Hello,
    Maybe off topic. I like the move. Harmony had a plan to go where Sun shunned in the past. Now Sun is forming a spearhead and want their implementation to still live foremost as a first choice in open-source only projects as well as many pre-exisiting ones.

    Giving the interest of the Harmony group, some may flock in as the largest group of non-Sun contributors. Also, at a point of completion, that open JDK will likely be included in some Linux-based distributions.

    The project will be big and maybe Eclipse and the Jazz collaboration framework (or a form of it) could help make this more at a stable and fast pace. Good day.

    Yemi Bedu

    Topic References:
    http://www.devx.com/Java/Article/28125
    http://blogs.zdnet.com/Burnette/?p=113

Leave a Reply

Your email address will not be published. Required fields are marked *