Ian Bicking: the old part of his blog

Re: Schevo and Durus

I'm trying to wrap my head around some of these issues also. I'm working on two projects now that both use RDBMS backends. One of them seems like a natural for using an OODB. It's a DB of network devices (routers, switches, APs, etc) and I've built my own simple ORM for it (could have used SQLobject, I know!). In many places, I already act like I'm using an OODB. E.g. instead of using crafted SQL to pull specific devices out of the DB, I loop over them all ('SELECT * from devices'), wrap each row up as an object instance, then check attributes on them to find matches (device.Model == 'Router M10'). I'll be examining this one to see how well something like durus could replace MySQL. I imagine it would be quite easy.

OTOH, the second DB is much larger and I think of as containing dumb data. A few well crafted indexes and queries are used to dump out search results as quickly as possible. There's nothing OO about it.

Comment on Schevo and Durus
by John P. Speno