Tony Morris

There are two kinds of logging and they have distinct answers: * reporting program states to users * unexpected program outcome, what was the program's previous state?

Not to sure what you mean by reporting problem states to users, is that API users? Also what do you mean by program state, how much information does that include?

The first was originally demonstrated in a Scala library called argonaut, which is providing program context ("logging") to API users, all the way to end users. By program state, I mean, "the program crashed, where was it up to?"

Sorry Argonaut looks like a json parsing library and not a logging library?

Correct it is, but it "logs" traversal of the zipper in a data structure. I could explain… not Twitter.

Thanks :) I am very interested in alternatives to the norm of ramming IDs everywhere.

I think it's most important to distinguish, "reporting program condition (e.g. errors) to users" and "oh no the program crashed, why?" Because they are handled completely differently.

Where can I learn more on finding out why a program crashed in a functional way?