Hello,

I also think that having a GDS interface to be implemented
with a Java-GDS and a JNI-GDS is the best approach.

I am going to define this GDS interface and post it
to this list in a day or two for your review.

I think that the teams needed would be:
1. A JDBC team that will implement the JDBC semantic
using the GDS interface.
2. A Java-GDS implementation team (at this moment I am
in this team and this implementation is a work in
progress) :-)
3. A JNI-GDS implementation team.

Any ideas ?

Bye.

Alejandro Alberola

--- In IB-Java@y..., "Ken Richard" <kenr@a...> wrote:
> Oops..wrong button. Let me try again.
>
> I do believe that this is the best approach, even just for the
purpose of
> getting development started. The JDBC semantics of the upper layers
of the
> driver are going to be the same no matter what mechanism is used to
talk to
> the server. I believe that the JNI stuff will be a lot easier to
implement
> and that we can get started there. I think we start by defining the
> internal interface betweent the JDBC semantics and the valves. Then
some we
> can divide up the work. Some can start working on the JDBC, some
can start
> working on a JNI-GDS valve and some can start working on a pure 3050/XDR
> protocol valve.
>
> The valve classes are implemented according to firebird/interbase
semantics
> and the jdbc classes are implemented according to jdbc semantics.
>
>
>
> -----Original Message-----
> From: news@n... [mailto:news@n...]On Behalf Of Mark
> O'Donohue
> Sent: Thursday, April 26, 2001 10:30 AM
> To: IB-Java@y...
> Subject: Re: [IB-Java] Re: JDBC Development
>
> <SNIP>
>
> Application
> -----
> DBXXXConnectionInterface
> -----
> with three seperate potential implementations (names for examples only)
> (DBXXXRemoteConnection,
> DBXXXLocalConnection,
> DBXXXInterserverConnection
>
> This in the jdbc client I would see acting much the same as Y-Value
> currenly
> does in the C client.
>
> The "LocalConnection" type 2 guy since it goes through the existing
C sql
> interface would additionally go through the C Y-Value layer to
determine if
> it was really local or connecitng remotely.
>
> <SNIP>