Re: character syntax fixes needed

[Dave Love]

Richard Stallman <address@hidden> writes:

> That is not the way we deal with standards in the GNU Project.
> We don't "ensure" that GNU Programs follow standards.
> We treat standards as suggestions, and we follow them when and if
> that seems like the best way to make GNU useful.

Of course it's useful, and I mentioned usage guides, not just
standards.  I assumed that was one reason to have the Unicode-based
Emacs you pressured me into doing a lot of work for on the basis that
no-one other than handa knew enough.

If you don't follow the relevant standards, GNU won't be interoperable
for a start, probably not even internally.  It will be a pain for
users if you don't follow their national standards or de facto
standards for things like character sets, keyboard layouts, fonts,
case conversion &c.  (For a relevant example, look at the mess with
XFree86/GTK/Mozilla internationalization, especially the inconsistent
handling of selections which has been biting Emacs.)  Fortunately much
of GNU now is consistently (I hope) based on Unicode and relevant
localization standards as it should be.

You also want to support open standards for political reasons, I'd
have thought.

> When Unicode gives us information about the usage of characters, that
> information can be useful for deciding what to do, and we should pay
> attention to it.  But this does not mean we necessarily follow
> Unicode's recommendations.  Following Unicode's recommendations might
> be a good default policy when we have no reason to do otherwise.

What on earth would be the reason to do otherwise, except what I said
about implementation and backwards compatibility?  Specifically, what
is the reason for the current treatment of guillemets?  It's clearly
incorrect, even going by the Oxford style guide I have on my shelf
which predates Unicode.  Whatever you think of some of its
implementation decisions, Unicode codifies a large amount of
linguistic information that appears to be trustworthy and I don't know
where else you'll find it collected.

> But Unicode's recommendations are not the final decision.

I'm not even talking about recommendations in this case, but the facts
about usage, and even what the Lisp manual says.  The net provides a
good way to confirm the use of quotation marks if you don't believe
the books.  If there are bugs in Unicode, you can get them fixed to
the benefit of all the GNU components that use its database.