Identifying resource-using streams

[Paul Sandoz]

On Apr 22, 2014, at 5:55 PM, Marko Topolnik <marko.topolnik at gmail.com> wrote:

> On 22. tra. 2014., at 12:41, Paul Sandoz <paul.sandoz at oracle.com> wrote:
> 
>> On Apr 21, 2014, at 10:54 PM, Marko Topolnik <marko.topolnik at gmail.com> wrote:
>> 
>>> With lazy streams, the scope of validity of a resource stops nicely following the scope of a block of code.
>> 
>> Yes, error handling is tricky, you may well want to react to various errors switching pipelines to return appropriate responses. It seems to me a more async/event-based stream model better fits your purpose above?
> 
> The Rx paradigm seems like a very good match, but I doubt there is a mature parallelizing implementation which I could use right now. Definitely worth investigating and keeping an eye on future developments. As far as I know, .Net already has a parallel implementation.
> 
> But more importantly, I see the need for the complete serverside stacks to support this processing model. This means that the entire stack must standardize on one approach and, naturally, I would first look to the JDK itself as the provider of that.


Or agree on the interaction between certain points. It's not clear to me yet whether a unified model is required.

It may well be possible to leverage an async/event model for some parts and an aggregated model for other parts. A tricky aspect here is working out the resource management balancing threads used for async tasks with those for computational divide/conquer tasks.

Paul.