Re: [Emacs-diffs] emacs-25 26b56dc: Fix some single quotes in documentation

[Paul Eggert]

On 03/16/2016 02:26 PM, Drew Adams wrote:
> > >They were used, and they weren't typos.
> Show us some that were not typos, please.

You want more examples? OK, here are a few in etc/NEWS.24:

   1919:**** Improved support for 'extern "C"' declarations in C/C++.
   1922:such as '__nonnull' and '__asm'.  Add '__cplusplus' macro when 
parsing C++.
   1931:**** Support for the Python 'WITH' and 'AT' keywords.

and in etc/NEWS before this week's changes:

     90:** The configure option '--with-gameuser' now allows you to 
specify a
     91:group instead of a user if its argument is prefixed by ':' (a 
colon).
     94:be installed setgid.  The option now defaults to the 'games' group.

There are more where these came from. These are not typos; they're the 
straight quoting style. Admittedly they are a minority of NEWS lines, 
but my impression is that the style's prevalence was growing in 
more-recent NEWS entries, even before this week's changes. And that's 
understandable, as the GNU coding standards in this area were changed 
relatively recently (compared to the entire Emacs timeline, that is).

> You wouldn't happen to know who started down
> that road, would you?

Markus Kuhn's web page on the topic 
<https://www.cl.cam.ac.uk/~mgk25/ucs/quotes.html> says that ISO 8859-1 
specified the current behavior where ` and ' are not symmetric; before 
that, ASCII was ambiguous and many displays used symmetric quotes. ISO 
8859-1 was derived from ECMA-94, which was developed sometime in the 
early 1980s by the ANSI X3L2 Technical Committee on Codes and Character 
Sets. So it appears that ANSI/X3L2 bears at least some of the 
responsibility for the road we're traveling on.

Personally I prefer symmetric quotes, and wish that ANSI/X3L2 had chosen 
otherwise. But we can't reasonably expect to change how Unicode works in 
this area.