> (though I can't remember how LO's URP implementations compute those globally unique TIDs; one would need to look into the code) I'd have to recheck too but I believe it uses a combination of the OS thread and process IDs. > Yeah, async messages complicate the picture somewhat. In olden times, UNO had a concept of [oneway] interface methods (that had to have a void return type, and don't do a response message over URP, and can be executed on top of another synchronous execution job that is currently calling out to another environment and is thus blocked waiting for a response). But today, the only remaining async operation in URP is the XInterface release operation Got it, and once thank you for the links. Best regards, Ritobroto