processi

about processes and engines

OpenWFE 1.7.0pre1

I’ve just released OpenWFE 1.7.0pre1.

I’ve taken away code I had written to flatten sequences. I had implemented it to allow for distributed sequences, i.e. sequences where each step is executed on a different engine (very much likely on another host).

Why this removal ? A good thing about blogging and exposing your ideas is that, while writing them down, it ignites new reflections based on older ones. I ended this ‘loop and cursor’ post talking about ‘flattening’ and I started to realise that there were other ways to assure distribution [of process segments]. It’s always a good thing to keep your algos simple and to remove adornments/cruft from them, ‘exceptional’ code is bad.

This removal ended up, as it ought to be, with extensive testing. I switched to an expression pool without caching and discovered 4 nasty persistence bugs, that I’ve just squashed.

Once again, I have to say thanks to everybody out there testing, using and giving feedback. Thanks !

Written by John Mettraux

March 21, 2006 at 2:01 pm

Posted in dev, openwfe, technical, workflow

%d bloggers like this: