April 14, 2010, 10:00 AM —
Having been inside a corporate acquisition or two, I can tell you that things are often very chaotic down in the trenches for the first few months or so. No one from the bought company really knows what the buying company really wants to do with them, and no one from the buying company knows how the acquired staff will fit in the new combined organization.
I get that, I really do. So I am fighting hard not to read too much in between the lines when yet another star Sun employee announces he is leaving Oracle.
This time it's James Gosling, the founder of Java, who blogged late last week that he had resigned from Oracle on April 2. This was after he had reassured Java developers in March that Java was going to be just fine under Oracle's new management.
Maybe Java will be okay, but it doesn't seem Gosling was terribly happy with Oracle. His blog skirts around the actual reason for his departure: "As to why I left, it's difficult to answer: Just about anything I could say that would be accurate and honest would do more harm than good." Coming from Gosling, who's well known for his colorful and direct nature, that kind of politic statement is alarming on its own.
What is also alarming is that Gosling is just the latest in a series of high-profile departures from Oracle by former Sun employees. February 3 marked former Sun CEO Jonathan Schwartz's last day, followed later that month by XML guru Tim Bray. Just a couple of weeks later, on March 8, Chief Open Source Officer Simon Phipps walked out the door. Gosling, it seems, held on for another month.
But curiosity abounds: under what circumstances? Is the new combined corporate atmosphere such a poisonous thing? To Oracle's credit, they may be just as surprised as we are. I have not met Gosling, but having spoken once or twice with Schwartz, Bray, and Phipps, I can tell you that their one common trait--being ridiculously hypersmart--might have put them at odds with the inevitable changes that such an acquisition brings. Then again, don't paint them all as daft technology wonks: I have personally seen Schwartz's and Phipps' business acumen in action, and I don't think this is a case of high-minded idealists walking out because the new guys are more business-savvy.
Even more worrisome, especially for open source watchers, is Oracle's continued non-approach to existing open source projects. I've already detailed the OpenSolaris situation on this blog, and elsewhere I have speculated that Oracle, rather than trying to kill OpenSolaris, may actually be moving it to a more traditional open-core model.
This week, Oracle has tried to allay fears about the fate of MySQL, when Oracle's Chief Corporate Architect Edward Screven announced at the O'Reilly MySQL conference Oracle's continued improvements in the beta release of MySQL 5.5, along with MySQL Cluster server 7.1, MySQL Enterprise Monitor 2.2, and SQL Workbench 5.2. (Note to Oracle: unified version numbers, please?)
A lot of people prior to the conference speculated that Oracle would hobble future MySQL development because why would it willingly invest in a database that would detract from its pre-existing product line? It turns out the answer is, because they want to go after customers Oracle's original products had a hard time reaching: Microsoft SQL Server users. According to ZDNet, Screven "also hinted at another key reason why Oracle bought the M in the LAMP stack: he noted that more customers deploy mySQL on Windows than on any other platform. That certainly gives Microsoft SQL Server a run for its money."
Indeed it would.
Also welcome was the news that the community edition of MySQL will still be supported. What is not entirely clear is just how many features will remain in the open source version of the database versus the proprietary enterprise version. Again, just like the current situation with OpenSolaris: there is a scarcity of information that is proving to be aggravating to open source developers. Oracle seems content to just let out trickles of information for now, as it tries to get things settled.
Some unsolicited advice to Oracle: don't wait too long to get your ducks in a row and start sharing your plans. Open source developers and users are historically wary of projects helmed by one big vendor. People didn't like some of Sun's shenanigans (read: control-freakiness), and it showed. The OpenOffice.org developer base shrank, MySQL development (when it belonged to Sun) was unfocused, and there was strong resistance to Project Indiana, the effort to create better package management and a distribution model for OpenSolaris. And, just to spread the love around, the open source community can get pretty testy with Red Hat, Novell, and Canonical, too, when they sense these vendors might be steering the ship too firmly.
That commercial versus community balance is ever-present in open source development. What Oracle needs to understand that, given enough frustration, customers and developers will start to vote with their feet. There's plenty of other forks in the sea, as well as completely separate open source projects and products for folks to code and consume, respectively.
Oracle may think itself immune to this kind of community walk-out, and maybe they are. But in the long run, whatever value they hoped to get from the acquisition of Sun would probably be lost.