Christopher J. 'CJ' Wright

You might also take a look at how this was done in https://github.com/xpdAcq/rapidz/blob/master/rapidz/clients.py and https://github.com/xpdAcq/rapidz/blob/master/rapidz/parallel.py

@dioptre rapidz was my friendly fork of streamz when I was using it to build software for x-ray sciences and needed a faster bug fix cycle. Rapidz also was built...

Both? I have realized that I have this pattern in some of my graphs (and I'm about to use it a bunch more) so my need for `gate` is solid,...

squash-and-merge works for me. I personally am often guilty of committing breaking code usually as a way to keep track of where I'm at in the middle of a refactor...

Do we want a `.github/contributing.md` to go along with this? (Can we just lift one off of another pydata project?) Edit: formatting

That's fair (maybe github should have an advanced contributing guide). Do you want to enshrine this information somewhere? If so, where?

I think this issue was really aimed at all those who have merging privileges, since we can do the squash via github and its our responsibility to make sure that...

This reminds me a little of the [Try-Monads talk from pygotham](https://2017.pygotham.org/talks/try-monads-with-big-data-using-pyspark/) (not that I think we should necessarily use Try-Mondads).

Relevent? http://reactivex.io/RxJava/2.x/javadoc/io/reactivex/exceptions/Exceptions.html

This might be one area where treating the graph as a walkable DAG may be helpful. One could imagine writing a function which walked the graph, looked for nodes where...