Hi Jörg,

On Feb 7, 2008 9:14 AM, Jörg Henne <j.henne@levigo.de> wrote:
Emmanuel Lecharny schrieb:
> Any other solution in mind ?
not having done any work on the backend makes it impossible for me to
make well-founded comments on the issue. However, the recent discussion
around splay trees and the problems with the current JDBM layer got me
thinking that the requirements of the database layer are almost entirely
identical to those of the RDBMS:
- entries are reasonably similar to tuples
- all indexes are essentially secondary indexes
- DNs are are similar to a primary key, but may also be treated like a
secondary one.

Yes exactly.  Also like an RDBMS we will eventually need to have some temp spaces for improving operations on certain search controls like sorting (i.e. building a special index for a specific search).  Just some ideas we've been tossing around.

Some not-so-lateral thinking brings up the question of whether it is
really necessary to invent another wheel. The idea may be totally silly,
and the required effort may easily make the think not worthwhile, but
what about having a look at Derby's store layer:

and in particular http://db.apache.org/derby/papers/btree_package.html

Good idea.  I did have these conversations a long long time ago with Daniel Debrunner while Derby was still incubating.  We found some issues but I think we could still share some code.  This is another option for us to consider.

Thanks this was an excellent option to point out.