MySQL Founder Monty Quits Sun (Or Not) 148
Paul Boutin writes "A reliable source tells Valleywag that MySQL inventor Michael Widenius, better known as Monty, has resigned from Sun. Sun bought Monty's MySQL company in a billion-dollar deal last January. Brian Aker, who forked the Web 2.0-friendly Drizzle SQL database (and former Slashdot engineer!), remains at Sun." Kaj Arnö and Sheeri Cabral share their thoughts.
Re:Well, I would do the same (Score:2, Informative)
Re:interesting timing (Score:3, Informative)
Actually, Maria isn't transactional. It was supposed to be, but it did't work out that way. Punting on transactions, they fell back to crash recoverable MyISAM. The next release is supposed to handle concurrent inserts. Other concurrent operations may follow someday. In the meantime, it doesn't support transaction backout, verb backout, two phase commit, transaction isolation, or any of the hard stuff.
MySQL went with InnoDB because MyISAM wasn't transactional. MySQL went with Falcon because Maria didn't have credible plan to become transactional. And that was almost three years ago.
Does Monty know how to write a transactional engine (or even think it's important)? Maybe, but so far, it's all smoke and no fire.
--JS
Re:Falcon architect Starkey also gone (Score:4, Informative)
Re:Well, I would do the same (Score:5, Informative)
Having slightly met Monty W, he is a true nerd. He didn't build a billion dollar company, he built a database that did what he thought databases should do. Many people do not exactly agree with him (see arguments on /. ad nauseam). But other people built a billion dollar company on that database. It deos not surprise me at all that he has taken his share of that billion dollars and walked off into the sunset. Maybe it is to Fiji, but even if it is, I would hazard a guess he will still be playing with databases on the beach.
Re:Chrome Users: new security flaw found (Score:3, Informative)
That's already been fixed in the Chromium codebase [google.com], r1677, but the latest download seems to be r1583. If I were the Chrome team, I'd be in more of a hurry to get this particular fix out there.