It doesn't seem that the RL-Glue protocol has been adopted by many (see here) and the project seems dead (see the 2018 update here), so I am unsure whether it's a good idea to adopt it, although I like the idea of providing a set of guidelines to define how agents should interact with environments and how to write the experiments code, in order to ease the interaction between different software (possibly) written in different languages. Maybe there's something newer than RL-Glue?
It doesn't seem that the RL-Glue protocol has been adopted by many (see here) and the project seems dead (see the 2018 update here), so I am unsure whether it's a good idea to adopt it, although I like the idea of providing a set of guidelines to define how agents should interact with environments and how to write the experiments code, in order to ease the interaction between different software (possibly) written in different languages. Maybe there's something newer than RL-Glue?