Re: *scratch* buffer documentation

["Óscar Fuentes" <ofv@wanadoo.es>]

Jean-Christophe Helary <jean.christophe.helary@traduction-libre.org>
writes:

>> I Think that you are missing one important thing: users are generally
>> not idiots.
>
> It is not about being idiots or not. But about how self-contained
> should the documentation be.
>
> If you show me the place in the documentation where the default
> behavior is described then I'll need to worry about why I did not find
> it.

  19.4 Killing Buffers

     ‘C-x k’ (‘kill-buffer’) kills one buffer, [...] If you ask to kill
  a file-visiting buffer that is modified, then you must confirm with
  ‘yes’ before the buffer is killed.


A corollary is that other buffers may not trigger the confirmation.

>> Emacs does have its dark corners, but I don't think saving buffers
>
> Indeed, saving buffers is not an issue. It is killing buffers that is.

IMO the implication that Emacs only cares about the persistence of
changes made to file-visiting buffers is clear. If the user, somehow,
expects that Emacs will warn on quitting about changes to *scratch*, he
will swiftly learn what "This buffer is for text that is not saved"
means :-)