On 31/07/07, Arnaud Simon <[EMAIL PROTECTED]> wrote:
> On Tue, 2007-07-31 at 10:45 +0100, Robert Greig wrote:
> > On 31/07/07, Rupert Smith <[EMAIL PROTECTED]> wrote:
> >
> > > I don't see any point in promoting a 'developer friendly' API in Java that
> > > is not JMS. The developer friendly messaging API for java is JMS.
> >
> > I completely agree.
> >
> > RG
>
> I don't agree as this interface should be used for exposing specific
> AMQP features.

Who are we to decide what specific features of AMQP will be useful to
expose. Either we create a low level API that exposes AMQP or we
don't? I thought that we had decided that a low level API was indeed
useful.

> The real debate is whether we wrap the AMQP classes
> around a higher level API. I personally think that we should as there is
> no notion of AMQP API and we should not introduce one.

The lack of a standard AMQP API should not be a barrier to Qpid
creating one if you really want to call it a Qpid API we can discuss
nomenclature once we know what the content is going to be.

If we don't start to work out what the API is who is? From what I
understand the AMQP-WG are busy working on the core protocol and not
standardising any API. This gives us the opportunity to provide
feedback to the group with a usable and tested API. If they decide
that it is not the direction they want to take that doesn't make our
work any less valid but if they do take it then we have an instantly
compliant client.


> So, we are thinking about defining a QPID QPI that may eventually impact upon 
> the
> JMS one.

The Qpid API should be impacting an AMQP API which in turn should be
impacting any JMS API changes. I don't think that a single
vendor's(Qpid) implementation would be strong enough to change JMS but
a new messaging standard (AMQP) would have the weight to make such
changes


-- 
Martin Ritchie

Reply via email to