Hi all, I really think we should work for terminology simplification but I wonder if it’s realistic in our timeframe. According to me the easy part is : - « Bean Class » for the class that gets Scanned : *Bean Class* (_quite obvious_) - « Bean Metadata for Bean<T> : *Bean Metadata* using « Bean » here is confusing) - « Required Type » for the type of injection type - « Session Bean » for EJB bean Now the more complex part is for the biggest one. From my understanding we have 2 concept with at least 3 different names : - Contextual Instance = Bean Instance = Bean - Contextual Reference = Client Proxy = Bean Reference I I’m not wrong on this classification, we should really choose an official name for these 2 concepts My personal choice would be « Bean Instance » and « Bean Reference » with perhaps a link to contextual names in chapter 6. The term « Bean » could be reserved to mean one to the other when difference is not relevant… What do you think ? Antoine Sabot-Durand ——————————————— Twitter : @antoine_sd CDI co-spec lead & eco-system development Agorava tech lead