> Do you have any long term (for FB 3.0, 4.0) plan to what new
> features will be implemented, what architectural changes will be
> done, as a big picture where do you want FB to be in the next 5
> years ?

Let's forget about versions, features can be assigned version after
they have been prioritized, sorted by dependency and checking the
available resources. So, let's concentrate on features. :)

> What is your "pet-project" for FB ?

1. External procedures, extending it to have callback notification
about transaction state changes (namely it would allow to
prepare/commit/rollback transaction against other database);

2. Full text search.

3. Slowly moving into direction of introducing structured data
types/domains, creating geoinformational types like "point",
"polyline", "polygone" and introducing geo-spatial index types and
queries.

3. Full XA compatibility.

The first one is moving in the right direction and can even make it
into FB 3.0 if we finally solve the interface issue. The second one is
terra incognita - none of the existing solutions are satisfactory
from my POV and none of the approached used in other databases is good
candidate for copying. The most hard part in the third solution is
adding structured types, indexing and querying part I already have in
Java with quite good performance. The last one is nice to have, but no
immediate actions are planned.

Roman