jaraco
commented
8 years ago That's a good point. Until very recently, that class was called "IRC" and instances of it were called "irclibobj", neither name which was helpful at all in describing the behavior.
I brainstormed alone and came up with Manifold because of the way it collected streams of data from different connections and handled them in a single place (the way a car manifold takes streams of air from the cylinders and combines them into a single exhaust).
I agree that Reactor would be better. Would you be willing to put together a pull request to include the changelog entry for a backward-incompatible release? Alternately, a blurb in the documentation may suffice.
seems to be global (?)
I'm not sure I follow. How is it global? What would prevent you from creating multiple Manifolds? I wouldn't recommend creating multiple, but I don't see anything that precludes doing so.
Original comment by: Jason R. Coombs
I've been going through the source-code for this library, and trying to write code that uses it.
From what I can tell, everything seems to be piped through the "Manifold" object, which seems to be global (?), and looks a lot like a reactor.
What the hell is a "Manifold"? It's not documented anywhere in the documentation, and where it's mentioned it's treated as calling it "the manifold" is self-explanitory.
The only results for googling lead to either the Manifold GIS system, or wikipedia pages for unrelated gometric terminology.
If it's functionally a reactor (and it certainly looks kind of like one), can you possibly use standard terminology, or at least add a blurb to the documentation ("the manifold is a reactor. We're using this non-standard term for {insert reason here}")?