On 14/03/2008, Carl Trieloff <[EMAIL PROTECTED]> wrote:
>
> It looks like M2.1 is landing nicely, so it might be time to start a
> thread on M3 scope. M3 work
> has begun as we all know but I think we need to block up what M3 will
> looks like, JIRA it and do the
> scope closing for it.

This all looks reasonable to me; my comment would be that we should
not make the scope too wide so that we can deliver a release that
gives real benefits in a reasonable time line, e.g. build complete by
June/July timeframe.

> This is my view of what M3 looks like -- please add, rant, comment etc...
>
> - Land the 'bigmergebranch' between M2.1 and truck
> - Tests, tests and more tests...
> - Java client M2.1(0-8/9) interop + AMQP 0-10
> - Java broker merged (still talking 0-8/9?)
> - Java broker (what else here??)
> - C++ Broker 0-10 (TCP +IB)
> - C++ Broker clustering module

Is clustering something that is likely to be ready for M3?

> - C++ Client 0-10
> - Ruby 0-10

Can someone give an update on the state of the Ruby module? Do we have
anyone who has the bandwidth to work on it? Is anyone using the
existing module?

> - .NET 0-10  (see rabbit has done WCP layer, do we want pick anything up
> from that as it is ASL, do we also want to support 0-8/9 from the same
> client, updating our client...)

I wonder if we want to continue working on our .NET client? I ask that
as a genuine question rather than a statement of my opinion. I think
it would be worthwhile understanding the limitations of a WCF approach
- people will not be surprised given my support of "extended JMS" in
Java to hear that a WCF (or extended WCF) implementation would to me
*seem* the best way of supporting our .NET users.

If the Rabbit client looks good then perhaps it makes sense to
recommend our users adopt it. I know we have struggled to find the
bandwidth to maintain our existing .NET client.

> Other items:
> - do we want to try have common events between brokers for mgmt?

As discussed elsewhere I think this would be good although I would not
think it a big issue if it was an M4 task.

> - I expect that we will work to get the Java broker both 0-8/9 and 0-10
> for M4

> Let's refine the list on mail and then build a JIRA set for M3.

In summary I think getting an M3 with the 0-10 protocol (backwards
compatible to 0-9) would be my focus, keeping the scope of the
released focussed on that.

RG

Reply via email to