For some reason

I'm excited by data... I just love to play with data, and I want to build a generic system to handle lots and lots of data.

I've built big systems using MySQL, and I generally find MySQL to be very good (I am looking forward to drizzle). However, RDBMS tend to get cumbersome and start to become annoying after time. There are five things I need out of a data store from a development stance.

1. Point-wise access: key-value pair
2. Index-And access: select * from table where f0=v AND .... AND fN=v
3. Transactions: Lock Row, munge data, Unlock Row
4. Index-And aggregation
5. Asynchronous replication and synchronization to another system

#1-#4 solve 95% of web-based persistence needs. The remaining 5% requires something special, and I've typically solve it with #5. This can be a giant pain in the ass.

I've taken various stabs over the years at trying to solve the problem, and basically they all kind of suck for various reasons. Now, I'm taking a different approach and starting a new open source project called: Nemo

Basically, Nemo is yet another key value pair system. More details are coming.