Always a hotly debated topic.

--jeremy

MySQL

Like Duh! MySQL!

I noticed all of the MySQL forks are on the list as well. I will throw my vote in for MariaDB, as solid MySQL "alternative".

How about OOo Base?

Base is a "desktop database management system" and is therefore not well suited for this category.

--jeremy

PostgreSQL, but I doubt it has a chance in hell of winning over the MySQL fanboys who've never even tried anything else.

It will be interesting to see what shakes out in the wash with MySQL now that it is owned by Sun, which in turn is now owned by Oracle. Likewise with innodb. I'm certainly not going to bet the farm on it seeing much "love" from Oracle. In any event, PostgreSQL is also far superior when you really need and actual RDBMS, acid, and some heavy lifting.

Oracle - partly because supporting it pays for my house, but mainly because they are huge linux boosters. Nice improvements in Oracle 11g as well, though I doubt I'll see anyone using them until Oracle 200g is out.

DB2 is also nice, though I find too few people use it.

Postgres. Not that MySQL is in peril lazy people have a reason to try it.

MongoDB is definitely worth watching in future if NoSQL/non-relational DBMSs are on the table.

I always watch those with interest, since I support non-RDBMS systems which are always considered 'inferior' to RDMS systems. It's actually rather amusing to see the same cycle repeating itself.

IMS and ISAM/VSAM are still just about the fastest and most scalable databases on earth, but most people who code with them would rather use SQL

I guess the world is not ready for a straight transition from a SQL scene to a query-less scene. Sure I'm not, not yet at least. So MondoDB sort of strike the chord with both parties involved. It could be an important link during the wide recognition of non-relational DBMS systems into mainstream.

I think you have that backwards - RDBMS came most recently. I honestly don't think that the transition to RDBMS makes any sense in a large scale setting, but then again I was one of the big fans of the technology when it started really gaining ground.

SQL took a world full of custom, non-portable query languages and methods (InnoDB, DBase, IMS, ISAM/VSAM, BerkelyDB, Info) and standardized on a single simple API.

That being said, RDBMS is also not a one size fits all, especially if all you're doing is reading data that's not truly relational or gains no benefits from normalization or de-normalization.

Probably a discussion for a different thread... but please bear in mind that this is not a new idea. This is an ancient idea that was retired for a reason, just like virtualization and clustering. There are places and times for every approach - and trotting out 40 year old techniques with a nifty new name does not invalidate the 40 years since they were considered cutting edge nor the reasons why they fell from popularity

PostgreSQL

I'm working in a high performance system,and use databases to share data, have persistent and high speed access; and MySQL works very fine.

MySQL