Hrvoje> Martin Buchholz <martin(a)xemacs.org&gt; writes: Hrvoje> This has been discussed a few months ago on comp.emacs.xemacs. Some Hrvoje> thoughts: Hrvoje> * It is logical for column-number-mode to count columns starting from Hrvoje> 1, because that is the way lines are counted by line-number-mode. Hrvoje> It was plain silly in older XEmacsen to press M-< and find yourself Hrvoje> on "line 1, column 0". Hrvoje> * what-cursor-position displays column numbers starting from 0 because Hrvoje> of the long history of doing so, and consistency with Hrvoje> `current-column' (the latter not being eligible to change because of Hrvoje> Lisp incompatibility the change would cause). When the Hrvoje> `column-number-start-at-one' feature was added, it was deemed "safe" Hrvoje> to change the relatively new modeline feature and leave the old Hrvoje> function alone. In a way, I approve of this conservative choice. Hrvoje> So column-number-mode needs to be consistent with two contradictory Hrvoje> things, which is impossible. Hrvoje> I think we should make `what-cursor-position' respect Hrvoje> `column-number-start-at-one' variable. Whatever we do, we have to be consistent. The user-visible and lisp-visible column-numbering have to match. Since (current-column) can't change, column-number-mode and what-cursor-position have to use 0-based column numbers as well (at least by default). If you add in Stallmacs-compatibility, the decision is clear. I vote for simply obsoleting column-number-start-at-one. I would be willing to accept the compromise of making it default to nil, but even that would be wrong. If we retain column-number-start-at-one, we would have to make what-cursor-position respect column-number-start-at-one, as Hrvoje suggests. I feel strongly about this. Isn't our current behavior obviously totally broken? It looks like a bug, so it is a bug. Martin