Splice out analysis? playback?

[jonahsmith]

Picking up from discussion here: #1 .

[jonahsmith]

Okay, so, the logic for splitting out the playback section would be that we modeled it as a generic mechanism, and what we have provided is actually more of an "example," correct? I can definitely see that argument, but realistically, I don't know we can ever expect there to be a real mechanism we could plug into.

I also want to raise another thought: splitting out the analysis code. I could see that having an afterlife in other projects.

What do others think?

[nikhilmitrax]

by analysis code you mean the one that takes the wav file and spits out decibels?

[jonahsmith]

Yeah, exactly.

(Edit to remove email forwarded junk.)

[nikhilmitrax]

well it's a thin wrapper around an open source library, so it should not really matter. But if you feel so, let me know. Since it's not necessary to run the code in any case, we could split it up.

[anothercookiecrumbles]

Don't think the point is as much that it's a thin wrapper. Instead, our project is meant to assume that there will always be a service that streams the data to us. In that sense, this code doesn't really belong here? I agree that we can never realistically assume that this source will exist, but if we split it up, we can simply add it as a dependency? I don't know; that just feels cleaner.

Maybe we can split out the whole 0_streaming_server, which has both: the playback and the analysis?

[nikhilmitrax]

we could just remove it from the repo... or do you guys want a separate repo for that?