Reza,
I'm now back from vacation and have had an opportunity to read your analysis and recommendation. I'd like to thank you
for the work you have put into this.
I think you have summarised very well both the strengths and weaknesses of providing such an API, and note your overall
conclusion that we should not consider the "utility API" approach further. I think my own views are similar.
(Any other views would be welcome)
In particular it seems to me that once we've extended it to offer similar features to the existing API we will have
something that isn't really any better, just different. We'll have simply replaced an OO API with a procedural API.
And, as you suggest, it lacks flexibility for all but the most trivial use cases. In particular, the "stateless" nature
of the API makes it more complicated to support "stateful" features such as transactions or auto-acknowledgement.
However I think you raise some very relevant issues about possible improvements to the basic JMS API, especially in the
area of how to instantiate and populate message objects. John's DI proposal raises the same issues. This is definitely
something we should look at further, and I'll make sure that we do.
Nigel
On 23/08/2011 01:25, Reza Rahman wrote:
> Nigel,
>
> Sorry for the delay in getting this to you. I was tied up with a few things and it look me longer than expected to get
> the analysis itself done. I figured getting this right the first time around is better than getting something out in a
> hurry. Here is the proposal: http://java.net/projects/jms-spec/pages/JMS2UtilityAPIProposal.
>
> I wound up not writing a proposal to make big changes to the basic JMS API itself. I really do not think it is warranted
> given the proposals we have at hand. Do let me know how you would like to proceed. If needed, we could have another
> conference call to discuss things in person.
>
> Cheers,
> Reza