Re: What makes an OODBMS different from an RDBMS?

Tim Mills wrote:
> > However we do already write data in the right order:
> > - Changed objects always get a new and free space in the database file
> > - write the object first
> > - switch the pointer afterwards
>
> This sounds like a shadow-object algorithm, similar to shadow-paging but at
> an object level. By what method do you ensure that allocation of free space
> and the pointer update is atomic and durable?

The pointer update currently is not atomic.

The planned algorithm:
The pointer "block" will consist of two pointers to the two versions of the respective object and one pointer to the transaction.

We keep track of free space (updated or deleted) with a tree mechanism, sorted by size and position. Adjacent freed areas are joined.

Are you working on something similar?
Do you need some more details? :-)

Kind regards,
Carl

---
Carl Rosenberger
db4o - database for objects - http://www.db4o.com