Seems like we might as well use the element name since that has to be unique; one less unique ID to store. Also if you have a schema, structs, etc. in your language of choice you now don't have to go update them for every possible element to add an id attribute, there aren't conflicts with anything that already has an id attribute that we inevitably will want to reference in the future, etc.
That being said, my concern is that some libraries would treat <subject/> as "{}subject" and some would treat it as "{jabber:client}subject". —Sam On Thu, Sep 9, 2021, at 13:15, Kevin Smith wrote: > It’s not the element name, it’s just (possibly unfortunately) the same > in that example. It’s the id. _______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: standards-unsubscr...@xmpp.org _______________________________________________