?

Log in

No account? Create an account

Where did all of the MySQL Developers Go?

« previous entry | next entry »
Mar. 22nd, 2010 | 10:25 am

As of late I have been getting a lot of questions around "Where did the MySQL developers go?'".

The answer is "elsewhere and nowhere".

The vast majority are still at Oracle, some have left, but plenty are still there. This got me thinking about "who wrote what". Innodb is a sizable piece of code and it continues to be at Oracle. Without Innodb, you don't really have a database that is 24/7. Innodb has been there for years.

What about the other pieces?

The optimizer guys are almost gone. They left a while ago to MontyAB.

Replication? They are all there minus the guy who wrote the first version (which is still pretty much what is used today). Him? He does consulting for different companies, but he left MySQL in 2002 or so. The execution/trigger system is a bit of combination. The "pre-stored procedures" stuff was all written back in the day, and little of it has changed, everything past that? That work was done by folks who are still at Oracle.

The parser? There is not much to the MySQL parser. A large chunk of it came from the merging of MiniSQL and Unireg (which is the only big thing I would say is wrong in Henrik's Map of MySQL Forks). Most of the additional work done there was for the monolithic stored procedure design that went into 5.0, and later stuff done by partitioning.

The guys who wrote the connectors are all at Sun still. I've not heard of a single one of them leaving.

The person who wrote stored procedures? He decided to go fishing back in 2005, I don't believe anyone has heard of him sense.

So the people who make up MySQL? Most are at Oracle.

How about the "group leaders" that made up the early MySQL leadership? Three are with MontyAB, another went fishing, and I work on Drizzle.

How about Drizzle? There are a few folks who did early patches on Drizzle that are still at Oracle (and I don't believe they are allowed to contribute any longer). Five are at Rackspace, one is at a storage engine company, and I am happily taking contracts on engineering work and spending the rest of my time playing with my aging dog. There are a couple more Drizzle folks who I suspect by the end of the year will be core developers and I am curious to see where they land (keep in mind, since our creation we passed 150+ contributors well over a year ago, so we have people all over the place).
Screen shot 2010-03-22 at 10.01.03 AM.png

Link | Leave a comment |

Comments {10}

Patches in and out

from: anonymous
date: Mar. 22nd, 2010 08:50 pm (UTC)
Link

Since I still have time before the user conference keynote, any additional info is of course welcome :-) About MiniSQL: MySQL doesn’t actually contain any MiniSQL code, does it? That’s certainly something I’m totally unaware of. (MiniSQL I of course know about, but it had already been forgotten when I became a MySQL user.)

Not that my map is focused on history anyway, but I like to get even that little corner right.

Btw, I thought things like statement based replication was also originally Monty's creation?

Henrik

Reply | Thread

Brian "Krow" Aker

Re: Patches in and out

from: krow
date: Mar. 22nd, 2010 08:54 pm (UTC)
Link

Hi!

MySQL certainly did at one point have MiniSQL code in it. I did a diff a while ago and there still is some, but more or less the entire thing has been written out by this point.

Statement based replication was written by Sasha. I spent a lot of time debugging it with him (along with Jeremy and others).

Most of the command line tools were pulled from this one guy's project off the internet (except for mysqlimport which was done by Jani and Monty).

Cheers,
-Brian

Reply | Parent | Thread

Arjen Lentz

Re: Patches in and out

from: arjen_lentz
date: Mar. 23rd, 2010 12:05 am (UTC)
Link

Which parts of mysql is that code?
Afaik it was only client/server stuff, but above you mention parser?

Reply | Parent | Thread

Brian "Krow" Aker

Re: Patches in and out

from: krow
date: Mar. 23rd, 2010 12:47 am (UTC)
Link

Talking to Arjen reminded me of a part of this... I know that Monty sat down with Hughes quite some time ago and Hughes had agreed that the COM bits were different enough that the work was not derivative (so... ask him for more specifics).

Reply | Parent | Thread