bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

This node mentions about keys following `C-c', but only `{', `}', `<',
`>', `:', `;', control chars, digits, letters, and "other punctuation".

What about keys `-', `+', `@', `#', `%', `^', `&', `*', `=', etc.?  These
are not mentioned.

I think, based on the mention of "other punctuation" coming just after
mention of the explicit list `{', `}', `<', `>', `:', `;',  that all such
chars are intended by "other punctuation".  I guess this because `<' and
`>', at least are not typically punctuation.

Could this doc please corrected to cover the cases of characters not
mentioned that are not really punctuation?  I'd propose that they should
be explicitly covered by what is said for "other punctuation": they
should be reserved for minor modes.

But regardless of what the decision is, even if it is to not put any
restriction on the use of the unmentioned chars, can we please make the
convention explicit in their regard, so there is no question, e.g.,
whether a sequence such as `C-c -' is reserved for minor modes?

Thx.

In GNU Emacs 24.5.1 (i686-pc-mingw32)
 of 2015-04-11 on LEG570
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --prefix=3D/c/usr --host=3Di686-pc-mingw32'

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Philipp Stephani]

[-- Attachment #1: Type: text/plain, Size: 1335 bytes --]

Drew Adams <drew.adams@oracle.com> schrieb am Di., 9. Feb. 2016 um
00:49 Uhr:

> This node mentions about keys following `C-c', but only `{', `}', `<',
> `>', `:', `;', control chars, digits, letters, and "other punctuation".
>
> What about keys `-', `+', `@', `#', `%', `^', `&', `*', `=', etc.?  These
> are not mentioned.
>
> I think, based on the mention of "other punctuation" coming just after
> mention of the explicit list `{', `}', `<', `>', `:', `;',  that all such
> chars are intended by "other punctuation".  I guess this because `<' and
> `>', at least are not typically punctuation.
>
> Could this doc please corrected to cover the cases of characters not
> mentioned that are not really punctuation?  I'd propose that they should
> be explicitly covered by what is said for "other punctuation": they
> should be reserved for minor modes.
>
> But regardless of what the decision is, even if it is to not put any
> restriction on the use of the unmentioned chars, can we please make the
> convention explicit in their regard, so there is no question, e.g.,
> whether a sequence such as `C-c -' is reserved for minor modes?
>

While there, could we also document the conventions for all other possible
key combinations? What about e.g. keys involving the Meta, Super, Hyper,
Alt, and Shift modifiers? What about C-F1 etc?

[-- Attachment #2: Type: text/html, Size: 1841 bytes --]

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

> While there, could we also document the conventions for all other
> possible key combinations? What about e.g. keys involving the Meta,
> Super, Hyper, Alt, and Shift modifiers? What about C-F1 etc?

This bug report is about C-c prefixed keys.  We have long had
binding conventions for C-c prefixed keys.

Please file another bug report if you want to suggest conventions
for additional keys.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Andreas Schwab]

Drew Adams <drew.adams@oracle.com> writes:

> This node mentions about keys following `C-c', but only `{', `}', `<',
> `>', `:', `;', control chars, digits, letters, and "other punctuation".
>
> What about keys `-', `+', `@', `#', `%', `^', `&', `*', `=', etc.?  These
> are not mentioned.

Aren't those covered by "other punctuation"?  It certainly matches what
C calls punctuation characters.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

> > This node mentions about keys following `C-c', but only `{', `}', `<',
> > `>', `:', `;', control chars, digits, letters, and "other punctuation".
> >
> > What about keys `-', `+', `@', `#', `%', `^', `&', `*', `=', etc.?  These
> > are not mentioned.
> 
> Aren't those covered by "other punctuation"?  It certainly matches what
> C calls punctuation characters.

As I said, I would like to read it that way also.  But there is no
mention of C or punctuation in a programming language, and those
are not usually considered punctuation in English etc., AFAIK.
And then there is the punctuation syntax class - does it mean that
these key conventions change, depending on which chars have
punctuation status in the current syntax table?

I would like the text to be a bit clearer about just what is meant,
including whether the notion of "punctuation" here is meant to
follow syntax class `.' (punctuation).

The bug report asks that we make clear which character keys we mean.
There may be other ones in other languages etc., but at least the
usual such keys on a typical US keyboard should be considered (and
so mentioned) as falling under what is said for "other punctuation".

For example, Lars recently proposed, in bug thread 22172, to bind
`C-c +' and `C-c -' by default.  Making the doc clearer about this
would preclude such a suggestion, as those keys would fall under
the category of "other punctuation", which is reserved for minor modes.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Andreas Schwab]

Drew Adams <drew.adams@oracle.com> writes:

> For example, Lars recently proposed, in bug thread 22172, to bind
> `C-c +' and `C-c -' by default.  Making the doc clearer about this
> would preclude such a suggestion, as those keys would fall under
> the category of "other punctuation", which is reserved for minor modes.

Then they cannot be used if we follow our conventions.  That looks clear
enough to me.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

> > For example, Lars recently proposed, in bug thread 22172, to bind
> > `C-c +' and `C-c -' by default.  Making the doc clearer about this
> > would preclude such a suggestion, as those keys would fall under
> > the category of "other punctuation", which is reserved for minor modes.
> 
> Then they cannot be used if we follow our conventions.  That looks clear
> enough to me.

You feel that the convention is clear enough for you.

Although I might agree with your interpretation of the convention
as currently stated, I would like it to be stated more clearly,
explicitly saying what it means by "punctuation", for example.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Andreas Schwab]

Drew Adams <drew.adams@oracle.com> writes:

> Although I might agree with your interpretation of the convention
> as currently stated, I would like it to be stated more clearly,
> explicitly saying what it means by "punctuation", for example.

It's just convention, use common sense.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Eli Zaretskii]

> Date: Tue, 9 Feb 2016 07:29:05 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: 22604@debbugs.gnu.org
> 
> > > For example, Lars recently proposed, in bug thread 22172, to bind
> > > `C-c +' and `C-c -' by default.  Making the doc clearer about this
> > > would preclude such a suggestion, as those keys would fall under
> > > the category of "other punctuation", which is reserved for minor modes.
> > 
> > Then they cannot be used if we follow our conventions.  That looks clear
> > enough to me.
> 
> You feel that the convention is clear enough for you.
> 
> Although I might agree with your interpretation of the convention
> as currently stated, I would like it to be stated more clearly,
> explicitly saying what it means by "punctuation", for example.

It's not really easy to define "punctuation" here.  More importantly,
some of the characters that _are_ reserved aren't punctuation by
Unicode categorization (they are symbols).

Would it help if we said "ASCII punctuation and symbol characters"?
That's the official name of those characters, AFAICT, and the "ASCII"
part makes sure no one will think about non-ASCII punctuation
characters, which I think we don't want to reserve.

Enumerating the characters can be tedious, so I think we'd like to
avoid that.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

> > > > For example, Lars recently proposed, in bug thread 22172, to bind
> > > > `C-c +' and `C-c -' by default.  Making the doc clearer about this
> > > > would preclude such a suggestion, as those keys would fall under
> > > > the category of "other punctuation", which is reserved for minor
> > > > modes.
> > >
> > > Then they cannot be used if we follow our conventions.  That looks clear
> > > enough to me.
> >
> > You feel that the convention is clear enough for you.
> >
> > Although I might agree with your interpretation of the convention
> > as currently stated, I would like it to be stated more clearly,
> > explicitly saying what it means by "punctuation", for example.
> 
> It's not really easy to define "punctuation" here.  More importantly,
> some of the characters that _are_ reserved aren't punctuation by
> Unicode categorization (they are symbols).
> 
> Would it help if we said "ASCII punctuation and symbol characters"?
> That's the official name of those characters, AFAICT, and the "ASCII"
> part makes sure no one will think about non-ASCII punctuation
> characters, which I think we don't want to reserve.
> 
> Enumerating the characters can be tedious, so I think we'd like to
> avoid that.

Yes, "punctuation and symbol characters" helps, IMO.  That was
one of my points: the chars are not necessarily punctuation, in
the usual sense of the word.

Dunno whether it's good to limit it to ASCII.  At a minimum it
should cover the "punctuation and symbol character" keys on a
typical US keyboard, of course.  Whether it should also cover
other punctuation or other symbol chars is maybe an open question.

It's really about character-inserting keys on a keyboard, not
punctuation and symbol chars that are not on keys.

But a key on a French keyboard that inserts a left guillemet
char should, I think, be handled the same by the convention as
is a key on a US keyboard that inserts a double-quote char or
a left angle-bracket char.

IOW, can/should we not apply the idea of such keys to keyboards
that insert other than ASCII chars?

I do agree that exhaustive enumeration would not be helpful.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Eli Zaretskii]

> From: Andreas Schwab <schwab@suse.de>
> Date: Tue, 09 Feb 2016 16:24:46 +0100
> Cc: 22604@debbugs.gnu.org
> 
> Drew Adams <drew.adams@oracle.com> writes:
> 
> > For example, Lars recently proposed, in bug thread 22172, to bind
> > `C-c +' and `C-c -' by default.  Making the doc clearer about this
> > would preclude such a suggestion, as those keys would fall under
> > the category of "other punctuation", which is reserved for minor modes.
> 
> Then they cannot be used if we follow our conventions.  That looks clear
> enough to me.

Indeed.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Eli Zaretskii]

> Date: Tue, 9 Feb 2016 09:24:03 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: schwab@suse.de, 22604@debbugs.gnu.org
> 
> Yes, "punctuation and symbol characters" helps, IMO.  That was
> one of my points: the chars are not necessarily punctuation, in
> the usual sense of the word.

Done.

> Dunno whether it's good to limit it to ASCII.  At a minimum it
> should cover the "punctuation and symbol character" keys on a
> typical US keyboard, of course.

Those are all ASCII, AFAIK.

> Whether it should also cover other punctuation or other symbol chars
> is maybe an open question.
> 
> It's really about character-inserting keys on a keyboard, not
> punctuation and symbol chars that are not on keys.

Not sure I agree.  Usurping too many characters would not be a good
idea, IMO.

> But a key on a French keyboard that inserts a left guillemet
> char should, I think, be handled the same by the convention as
> is a key on a US keyboard that inserts a double-quote char or
> a left angle-bracket char.

How can such a binding be useful, when Emacs runs on a non-French
keyboard?  Do we really want to encourage modes that can only be fully
functional in some specific locales?  I don't think so.

Any reasons not to close the bug report?

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Drew Adams]

> > Yes, "punctuation and symbol characters" helps, IMO.  That was
> > one of my points: the chars are not necessarily punctuation, in
> > the usual sense of the word.
> 
> Done.

Thx.

> > Dunno whether it's good to limit it to ASCII.  At a minimum it
> > should cover the "punctuation and symbol character" keys on a
> > typical US keyboard, of course.
> 
> Those are all ASCII, AFAIK.

Yes, I think so.

> > Whether it should also cover other punctuation or other symbol chars
> > is maybe an open question.
> >
> > It's really about character-inserting keys on a keyboard, not
> > punctuation and symbol chars that are not on keys.
> 
> Not sure I agree.  Usurping too many characters would not be a good
> idea, IMO.

Usurping?  From whom/what?  It's about reserving this category
of keys for minor mode keymaps.  The question is which punctuation
and symbol chars?  And the answer, I think, is those that are
associated with keyboard keys (one key inserts one char).
For a US keyboard that means only ASCII punctuation and symbol
chars, but for other keyboards it can mean other chars.

> > But a key on a French keyboard that inserts a left guillemet
> > char should, I think, be handled the same by the convention as
> > is a key on a US keyboard that inserts a double-quote char or
> > a left angle-bracket char.
> 
> How can such a binding be useful, when Emacs runs on a non-French
> keyboard?  Do we really want to encourage modes that can only be fully
> functional in some specific locales?  I don't think so.

Why treat punctuation and symbols on a US keyboard specially?
Why should `$' be treated differently from `€'?

> Any reasons not to close the bug report?

You can close it if you like.  I would prefer that we think about
handling the same category of char-inserting keys (punctuation and
symbols) for other keyboards the same way we handle those for US
keyboards.  But if that is not to be then it is not to be.

Thanks for working on this.

bug#22604: 24.5; (elisp) `Key Binding Conventions': what about other `C-c' keys?

[Eli Zaretskii]

> Date: Tue, 9 Feb 2016 10:33:28 -0800 (PST)
> From: Drew Adams <drew.adams@oracle.com>
> Cc: schwab@suse.de, 22604@debbugs.gnu.org
> 
> > > Whether it should also cover other punctuation or other symbol chars
> > > is maybe an open question.
> > >
> > > It's really about character-inserting keys on a keyboard, not
> > > punctuation and symbol chars that are not on keys.
> > 
> > Not sure I agree.  Usurping too many characters would not be a good
> > idea, IMO.
> 
> Usurping?  From whom/what?

From users and from other modes.

> It's about reserving this category of keys for minor mode keymaps.

Yes, and I don't think we want or should reserve too many characters.

> > How can such a binding be useful, when Emacs runs on a non-French
> > keyboard?  Do we really want to encourage modes that can only be fully
> > functional in some specific locales?  I don't think so.
> 
> Why treat punctuation and symbols on a US keyboard specially?
> Why should `$' be treated differently from `€'?

Because most keyboards have the former, but not the latter.

> > Any reasons not to close the bug report?
> 
> You can close it if you like.  I would prefer that we think about
> handling the same category of char-inserting keys (punctuation and
> symbols) for other keyboards the same way we handle those for US
> keyboards.  But if that is not to be then it is not to be.

We can think with the bug closed, and we can reopen later (or make a
change without reopening).