Kyle Jones <kyle_jones(a)wonderworks.com&gt; writes in xemacs-beta(a)xemacs.org: ... Yes. We apparently agree to disagree on issues like this. Signaling errors when attempting to move beyond buffer boundaries is idiotic behavior, but enshrined for all-time as "standard". I've been overruled on this issue, but that doesn't stop from attacking other windmills. Setting the mark on these movement commands is different than all the other movement commands which don't. It's inconsistent and stupid (IMO). I'd suggest coming up with new functions for `beginning-of-buffer' and `end-of-buffer' which do not set the mark and allowing users to bind them to M-> and M-< if they see fit.