kwalker at pentaho.com
Fri Feb 1 09:23:48 EST 2013
On Thursday, January 31, 2013 08:36:50 PM Julian Hyde wrote:
> On Jan 31, 2013, at 12:35 PM, Kurtis Walker <kwalker at pentaho.com> wrote:
> > It looks like Mondrian has a home grown implementation of the Actor pattern. It does not use any third party libraries for Actor implementations. Is that correct? I'm curious why we decided to roll our own instead of using a library off the shelf.
> That's my fault. I decided to implement an actor mechanism from the ground up (or rather, from threads and queues) when we introduced the SegmentCacheIndex. Actors are more of a pattern rather than an algorithm, so I wanted to be able to refactor the existing code without having to fit into the model imposed by a framework.
> I looked at Akka at the time, and the problems it solved weren't the problems I was worrying most about (all of which were about our own algorithms and data structures). If you think there would be a benefit to moving to an actor framework, I am open to it.
I don't know that there would be a benefit at this time. But, it's something to keep in mind if we write more features that use them.
> Mondrian mailing list
> Mondrian at pentaho.org
More information about the Mondrian