Hi Timothy
I think that, with regard to the new SE interface and the reader/writer
paths, things are a bit too barebones to split that work, at least for the
first SE implementation.
I mean once there's at least one SE implemented and things are
relatively defined it'd be possible, but right now we'll just be working on the
same code, which will lead to merge frustration.
I have started to work on the read path of the File SE (read path is
pretty much the one that interests me for now) but I don't *have* to work on it.
I'd actually prefer to work on any other portion that is in the
critical path for distributed execution.
Jacques: wdyt? what are you focusing on and what can I take off your
hands?
You mentioned you were focusing on parsing/optimizing, maybe I could
try and start to connect the dots and add tests somewhere else?suggestions?
Best
David
On Apr 19, 2013, at 3:58 PM, Timothy Chen <[email protected]> wrote:
> Hey David,
>
> Just chatted with Jacques earlier and I decided to punt on the Supersonic
> stuff for now, and help on the framework that we have.
>
> I originally wrote the ScanJson a while back when we didn't have anything
> setup, and now we do have the interfaces I'll like to get it working again.
>
> Jacques told me you're working on the Storage Engine, and proposed that we
> can work together on that where we can start from the File based SE and I can
> start on the read path while you do the write path.
>
> Let me know what you think,
>
> Tim
>
