trex-core traffic generator question (STF vs. STL vs. ASTF)

Trying to get t-rex up and running (its been a real pain).

I am looking for some clarification on the following. I really don't understand the difference between Stateful (STF), Stateless (STL), and Advanced stateful (ASTF) modes. After reading the documentation this is what I gather, can someone clarify if I am correct or wrong?

• STF: This is old, not recommended to use anymore, but mainly you tell trex to create a stream. One stream at a time. Configured with .yaml files to work with the API.

• STL: This is the preferred method to use (I think?). Allows TREX to work as a server. Run the system in interactive mode and people can connect with the stateless gui, or console, or whatever means to send multiple streams.

• ASTF: From what I understand this can be run in interactive mode (./t-rex-64 -i --astf), and mainly does the same thing as STL, but uses python files to interact with the API.

I read though the documentation here, its still not clear to me what the benefit that ASTF provides? The name is just misleading to me, as it seems to be performing stateless operations as t-rex can be run in an interactive fashion.

I find trex-core documentation has lots of stale entries of files that no longer exist in the repo, or just lacking any clarity. I feel there should be some table that compares all three of these modes as its not very clear why I would want to use ASTF over just STL...