Clarify In/Out terminology

[alexcrichton]

Commented here: #62 (comment)

I personally interpret in/out with one of two interpretations:

• Read out a request, write in a response.
• Read in a request, write out a response.

Unfortunately these mean opposite things :(. Looks like this is standard terminology in finagle though? (cc #62 (comment)). Maybe I should just shift how I view these?

(may also be worth exploring alternative names)

[seanmonstar]

FWIW, in hyper I have the concept of Incoming vs Outgoing. This distinction appears in other HTTP libraries that I've looked at. This causes a very odd-looking case:

impl<T> FramedIo for Conn<T> {
    type In = T::Outgoing;
    type Out = T::Incoming;
    // ...
}

[carllerche]

In my terminology, In = message going into the socket, Out = "message coming out of the socket"

What is incoming / outgoing in your case?

[seanmonstar]

Incoming = message incoming to this machine, Outgoing = message going out from this machine.

[tikue]

This is confusing every time for me. I don't think inverting the in/out cases would make it any more intuitive.

[aajtodd]

Usually seen as the second choice: Read in a request, write out a response. I think it would be very confusing the other way around.

This matches Finagle Service definitions Finagle Services as well as Wangle (and pretty much any socket library I've worked with for what it's worth).

What I don't see or haven't wrapped my head around is the filters part of Framing for tokio. I know I've gone back and forth with @carllerche on this in #63 and gitter but embedding the types that way seems rather limiting and requires lots of boilerplate and manual plumbing.

[carllerche]

There is no more In / Out.

[tikue]

There is in Codec. Is it worth reopening this in tokio-core, or is it less ambiguous in that usage?

[alexcrichton]

@tikue sure yeah, I've opened one: tokio-rs/tokio-core#135