Open source and management disputes

Open source and management disputes

Summary: Marc spoke out of turn, saying he wasn't happy with the R&D investment he was getting from home office. In a traditional company, a division manager who speaks out of turn like this gets tossed out on their ear. His successor learns to keep his head down and his mouth shut. But to do that here would be a huge mistake.

TOPICS: Open Source

Open source vendors can't be run like their closed source counterparts. The programming talent, and the people behind that talent, aren't like other managers.

I thought of this when looking at a stray comment Marc Fleury of RedHat's JBOSS unit gave another publisher recently. They ran with it like sports reporters ran with Michael Vick giving fans the finger.

Marc is not a traditional manager. He's an entrepreneur, a programmer. He's more like the people you'd find in the music or movie industry than the software industry, only with less guile. He's a lot more like Vick than like Falcons coach Jim Mora.

Marc spoke out of turn, saying he wasn't happy with the R&D investment he was getting from home office. In a traditional company, a division manager who speaks out of turn like this gets tossed out on their ear. His successor learns to keep his head down and his mouth shut.

But to do that here would be a huge mistake. It would cause enormous turmoil among employees and customers. Fleury could be out building a new company the next day, perhaps in a competitive niche.

The only thing protecting RedHat from direct competition would be contract language. Fleury's knowledge of the codebase could be re-used as soon as the contract's non-compete clause ran out. And all the talent he brought with him could leave as well.

This is also a tough moment for RedHat. The threats of Oracle, Novell and Microsoft are real, massive. Fleury also has a good point. JBOSS faces new threats from IBM and the best way for RedHat to secure its future is to invest "up the stack" in middleware and enterprise applications.

This has to be handled gently. RedHat shouldn't ask what Microsoft or IBM or Oracle would do. It needs to find an open source way. 

I will be very interested to know what that way is.

Topic: Open Source

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • "He wasn't happy with the R&D investment..."

    "... he was getting from home office."

    Yes, open source does have unique versions of ordinary problems. Two applicable characteristics of open source are reliance on reduced staff costs as a means to profitability and restrictions on profit as a result of unpaid availability of the product.

    In the circumstances of a public company, how is extra R&D investment to be justified acceptably to stockholders? Even Microsoft lost ground when the company announced a huge investment in the resources needed to enter large new markets.

    If Red Hat is required to compete on the basis of a stack, how is development of the stack to be funded?
    Anton Philidor
    • How Development is funded

      Open source has two sources of funding. Internal sources from whoever is managing the process, and external efforts from the community.

      It's a fair criticism of Fleury for RedHat to say he hasn't engaged his own community enough, and perhaps to move in someone as "community manager" who will stimulate greater efforts.

      These efforts aren't just code. They're also requirements, and priorities, as well as testing services. Organizing and maximizing those efforts is important.

      Is JBOSS doing enough in this area? I don't know. I'd love to hear more from those inside that process.
      • Merging cultures.

        Apparently JBoss is holding to the internal funding model in use at time of purchase. What struck me about the quote was the absence of acknowledgement of any change in approach.

        If Red Hat decided that the JBoss approach will have to change, and can sell that to its new staff, how long will be required to obtain a sufficient community? Particularly given your observation that Red Hat's existing services have a problem with sufficient engagement.

        You're right about needing to know more about what is occurring.

        Thinking further, if Red Hat expects to have a stack to sell, how many cultures will have to be changed to obtain it?

        That's a broad strategic question about what software, how to integrate it for Red Hat's stack, and how to make certain it remains responsive to Red Hat's goals.

        That may be essential to Red Hat's survival. The JBoss problem may be a model for what Red Hat will face in the future...
        Anton Philidor
      • What JBoss community?

        Dana, what JBoss community could provide the 'external efforts' that you mention? <br /><br />

        We all know, as your readers have mentioned here, that JBoss is largely developed by folks who have an or email address. <br /><br />

        This is likely the #1 reason that Geronimo has been able to close the gap to JBoss so quickly...i.e. the Geronimo community actually is a community with over 20 companies with committers. <br /><br />

        PS: If Marc wanted R&D $$$$ he should have gone to the prom with Oracle. They wouldn't have cared that JBoss doesn't have a community... <br /><br />

        Savio<br />
  • The Red Hat and the JBoss models are very different.

    With Red Hat, they help with the Linux Kernel, but are a small part of the total effort. They make money by giving support and being able to guarantee they can fix customer problems.

    With JBoss, almost all of the developers are in house and it is almost completely developed from within. This makes it a lot harder to pay for development out of the support dollars.

    What Red Hat really needs to do is find ways to get others to share the development, and have a common code base. Right now, we have too many open source and comercial projects that do the same thing JBoss. The efforts among the Open Source groups at least need to be unified.
    • Not after Red Hat bought JBoss.

      Perhaps development funding is so different that the Red Hat approach cannot be applied to JBoss. If that's true, then Red Hat made an expensive mistake.

      And how much of the "stack" Red Hat hopes to accumulate uses the JBoss model?
      Anton Philidor
      • Well, JBoss is still developed almost all in-house, and Red Hat Linux is

        the cooperation of many, Red Hat being a very small part or the total development.

        So, with JBoss, they have to charge enough for support to pay for all of the development, and of course the support.

        Yes, Red Hat may have been smarter to support Gernonimo for instance.
      • Costly mistake indeed

        Anton, <br /><br />

        I suspect that Red Hat made the mistake of thinking that because JBoss is under the GPL, that means there is a vibrant and diverse community around JBoss. An understandable mistake to make considering that this would be true in the Linux community. If only Red Hat had paid attention to the JBoss committer list or the mailing list...<br /><br />

        Pretty much all of the 'stack' that Red Hat is going after is being delivered via JBoss, and their very closed-nature of open source development. It will be interesting to see if JBoss gets more R&D cash, or Red Hat continues to spend it on SG&A.

        • The right way for RedHat to proceed

          Changing the subject from what Fleury said, or what he said he wanted, to making the JBOSS community more open and vibrant may be the right way for RedHat to launch its discussion of this internally.
          • Community building isn't easy...

            If only it were that easy to build a community...I personally believe that just having an open source product does not get you anywhere close to building a community.<br /><br />

            I can't think of a vibrant & open community that is lead by a commercial entity. Eclipse, Linux, Tomcat, Apache HTTP, Firefox, Geronimo, etc. are all examples of open source projects with vibrant & open communities without a commercial entity resembling the project lead. This doesn't mean that open source projects without such stellar communities can't succeed......just that the cost of development isn't shared across multiple individuals/vendors/groups.<br /><br />

  <br /><br />

            Savio<br /><br />