Good points Roman; I hadn't considered the /etc/default/hadoop* when filing this.
Indeed that acts in ways similar to hadoop-env.sh (does it also allow heap size control, etc., which is primarily what people modify it for?), but its not very discoverable.
Having an empty file is worth it (and is what I filed for as well, sorry I did not elaborate early) - 'hadoop-env.sh' is written about in several guides online as a place to set stuff like specific daemon heap size, etc.; which a new Bigtop package user may easily miss?