Red Hats acquisition of JBoss, like any good story, offers interesting views from at least three angles.
The first angle is what it says about the industry. The second angle is what it says about the technology. The third angle is what it says about the craft of being a software development professional.
What this deal says to would-be entrepreneurs, concerning the state of the software industry, is that the dream is not dead—but it is different. The late-1990s dream of starting your own company, growing it on a fistful of maxed-out credit cards, and taking it public to become a self-made millionaire has bumped up hard against the realities of operating a public company today.
An April 10 blog posting from JBoss honcho Marc Fleury says it all: “… an IPO was possible and indeed planned but in this day and age of Sarbannes-Oxley[sic] the right M&A provides liquidity and reduces much of the risk.”
Translation: Taking the company public means spending more time with lawyers than with the “60 developers and 30 absolute superstars” that Fleury is proud to employ, as he recently told BusinessWeek.
Coincidentally, members of eWEEKs Corporate Partner advisory board told me something quite similar in a conference call April 10 on the subject of SOX and other mandates: Theyre definitely spending more time than they used to on the aspects of IT function that have little direct connection with technology, but represent critical concerns about documenting processes and satisfying regulatory burdens.
To refract this issue at another angle, this means that theres entrepreneurial opportunity in giving those enterprise technologists more unified tools for monitoring and managing their stacks from the viewpoint of “am I in compliance today?”—rather than just “are my systems up and running?” The latter continues to be important, but taking a system down—before its noncompliance gets you in trouble—may be part of the mission of next-generation IT management, as opposed to declaring victory in proportion to the negative of the logarithm of ones downtime fraction.
What this deal says about technology is that integrated stacks have a commanding edge in performing the tasks that matter most today—for example, to tasks like assuring SOX compliance. An operating system cant know the difference between a business process and a low-level input/output task, but a middleware layer can know that difference—and when the OS and the middleware become a well-integrated intelligent platform, good things can happen, as Microsoft may someday demonstrate if it brings something like LINQ or WinFS to market.
For the professional software developer, this story should resolve any lingering doubts about whether open-source development is professional software development. Clearly, this is not a leisure-time activity pursued by people getting full-time salaries for less than full-time jobs. I wouldnt try to draw a line and say when that became clear in my own mind, but its been several years since I first advanced the idea that software development is more like being a lawyer than being a plumber.
The latter only gets paid for one pipe joint at a time, but a lawyer gets paid to do two things: first, to add to the body of knowledge thats used by everyone—including competitors—by participating in cases that lead to decisions available to all; second, to apply personal familiarity with that body of knowledge in the interest of particular clients. Its not the rock-star living that comes from writing Windows and getting paid hundreds of dollars for each incremental copy that has an actual cost of about $1. It is, however, a craft in which one can take some pride.