Re: transpose-chars, transpose-words at the beginning and end of buffer

[Eli Zaretskii]

> From: Dan Jacobson <jidanni@kimo.FiXcomTHiS.tw>
> Newsgroups: gnu.emacs.bug
> Date: 05 May 2001 02:28:59 +0800
> 
> What a shock it was to find there are still some situations where C-t
> doesn't do the "right thing" and instead flashes the screen: when one
> is on the first character of the buffer.  I mean it already does "If
> no argument and at end of line, the previous two chars are exchanged."
> so why not sugar it up some more? [want 1234 to become 2134].
> 
> Problems: but then should the current behaviour when at the beginning
> of a line also be redefined?  What about transpose-words?

I don't think the behavior of these two functions must be consistent
if a less consistent behavior makes more sense to the users.

C-t does what it does at the end of a buffer because that situation
frequently happens while typing (with people who type with both
hands).  Typing two words in the wrong order is much less frequent.

> Wait a second: transpose-words already "does the right thing" when at
> the begging of buffer!  However it this special behaviour is not
> documented in Help or Info.

That's because it's not special: if point is at the beginning of the
buffer, the first word is still ``the word before or around point''.