Let's not rush to (architectural) conclusions

- We all agree that external processors are a good idea. This is
progress from 2 years ago. Personaly I have doubts about Perl, PHP
and alike, but who am I to judge somebody else's needs?

- It is also new, sort of. Let's take the time to do some plugins and
figure out the pro's and con's, the issues, etc. See debate about
where to stick the code for example. Nothing like real life use to
get a feel for what works best. In the meantime, the Fyracle build
can act as testbed. The Firebird side of it is fully open source,
IDPL.

- In my view the isc_* vs. JDBC engine API is a long term issue.
We'll need the isc_* API for a while because all existing language
drivers are written against that API. It will take time to change
that. Moreover, one is a thin wrapper around the other, and who wraps
whom is not an urgent decision, IMHO. No need to "fight it out" now.

- I don't actually see why layering is broken, now that SQL has moved
into the engine. Sure, FB10/FB15/FB2 break layering, but that is a
different debate -- and I think we all agree now that layering should
be restored. Nothing to see here anymore, let's move on.

Paul