TUTORIAL.de updated

TUTORIAL.de updated

[Werner LEMBERG]

I've done a quick update of TUTORIAL.de -- there are certainly issues
which can be improved, but I'm short of time, unfortunately.

It's revno 106888.


    Werner

Re: TUTORIAL.de updated

[Martin]

Werner LEMBERG writes:

> I've done a quick update of TUTORIAL.de -- there are certainly issues
> which can be improved, but I'm short of time, unfortunately.
>
> It's revno 106888.
>
>
>     Werner

Details ?

Re: TUTORIAL.de updated

[Werner LEMBERG]

>> I've done a quick update of TUTORIAL.de -- there are certainly
>> issues which can be improved, but I'm short of time, unfortunately.
>>
>> It's revno 106888.
>
> Details ?

The section `MULE' (which doesn't exist in TUTORIAL) could certainly
be improved.


    Werner

Re: TUTORIAL.de updated

[Eli Zaretskii]

> Date: Thu, 19 Jan 2012 10:37:12 +0100 (CET)
> From: Werner LEMBERG <wl@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> The section `MULE' (which doesn't exist in TUTORIAL) could certainly
> be improved.

Or, rather, removed?  Translations of the tutorial are not supposed to
have text that doesn't exist in the English tutorial.  If we allow
that, we will have a tremendous maintenance problem on our hands,
whereby maintainers cannot be sure they control what is or isn't in
the package.

If a translator wants to suggest additions, she should do it first in
the English tutorial.

Re: TUTORIAL.de updated

[Werner LEMBERG]

>> The section `MULE' (which doesn't exist in TUTORIAL) could
>> certainly be improved.
>
> Or, rather, removed?  Translations of the tutorial are not supposed
> to have text that doesn't exist in the English tutorial.

Hmm.

> If we allow that, we will have a tremendous maintenance problem on
> our hands, whereby maintainers cannot be sure they control what is
> or isn't in the package.

The section how to select a German input method, and how to select
Latin-1 encoding (and now UTF-8) was there from the very beginning of
the translation.

> If a translator wants to suggest additions, she should do it first
> in the English tutorial.

Well, a typical US user doesn't need to know how to enter foreign
languages, does she?


    Werner

Re: TUTORIAL.de updated

[Stephen J. Turnbull]

Werner LEMBERG writes:

 > > If a translator wants to suggest additions, she should do it first
 > > in the English tutorial.
 > 
 > Well, a typical US user doesn't need to know how to enter foreign
 > languages, does she?

No, but a translator to a language that doesn't yet have a version of
TUTORIAL will most likely start from English.

Re: TUTORIAL.de updated

[Werner LEMBERG]

>  > > If a translator wants to suggest additions, she should do it
>  > > first in the English tutorial.
>  > 
>  > Well, a typical US user doesn't need to know how to enter foreign
>  > languages, does she?
> 
> No, but a translator to a language that doesn't yet have a version of
> TUTORIAL will most likely start from English.

Ah, yes, I haven't thought of this.  Unfortunately, I don't have time
to translate the MULE section to English, sorry.


    Werner

Re: TUTORIAL.de updated

[Eli Zaretskii]

> Date: Thu, 19 Jan 2012 14:02:03 +0100 (CET)
> Cc: parozusa@web.de, emacs-devel@gnu.org
> From: Werner LEMBERG <wl@gnu.org>
> 
> The section how to select a German input method, and how to select
> Latin-1 encoding (and now UTF-8) was there from the very beginning of
> the translation.
> 
> > If a translator wants to suggest additions, she should do it first
> > in the English tutorial.
> 
> Well, a typical US user doesn't need to know how to enter foreign
> languages, does she?

Probably not, but then why does a German user need a lesson on how to
type German on her keyboard?  Is a situation where a native German
speaker learns to use Emacs on a machine that doesn't already support
German input on the OS level anything but very rare?

Re: TUTORIAL.de updated

[Werner LEMBERG]

>> Well, a typical US user doesn't need to know how to enter foreign
>> languages, does she?
> 
> Probably not, but then why does a German user need a lesson on how
> to type German on her keyboard?

It is not too uncommon that German users have a US keyboard because
they are programming a lot.  The characters [ ] { } are really awkward
to type on a German keyboard.

> Is a situation where a native German speaker learns to use Emacs on
> a machine that doesn't already support German input on the OS level
> anything but very rare?

Well, on my GNU/Linux box, the Quail input methods for German offered
by Emacs are better than SCIM.  One main reason is that the latter
disables the use of a `.XCompose' file, which is very bad for me.  I
don't know whether the successor of SCIM (I forgot its name) has fixed
this.

Regardless of the keyboard and input method issues, there are still
three dominant encodings used in Germany and Austria, namely latin-1,
windows-1252, and utf-8, and I consider it important that a user knows
how to force an encoding in case Emacs fails to recognize it properly.

As I said, the MULE section can be certainly improved :-)


    Werner

Re: TUTORIAL.de updated

[Eli Zaretskii]

> Date: Fri, 20 Jan 2012 06:00:22 +0100 (CET)
> Cc: parozusa@web.de, emacs-devel@gnu.org
> From: Werner LEMBERG <wl@gnu.org>
> 
> 
> >> Well, a typical US user doesn't need to know how to enter foreign
> >> languages, does she?
> > 
> > Probably not, but then why does a German user need a lesson on how
> > to type German on her keyboard?
> 
> It is not too uncommon that German users have a US keyboard because
> they are programming a lot.  The characters [ ] { } are really awkward
> to type on a German keyboard.

Those awkward characters don't appear in the tutorial, and we
certainly don't ask the apprentice to type them as part of her working
on the tutorial lessons.

> > Is a situation where a native German speaker learns to use Emacs on
> > a machine that doesn't already support German input on the OS level
> > anything but very rare?
> 
> Well, on my GNU/Linux box, the Quail input methods for German offered
> by Emacs are better than SCIM.  One main reason is that the latter
> disables the use of a `.XCompose' file, which is very bad for me.  I
> don't know whether the successor of SCIM (I forgot its name) has fixed
> this.

Emacs makes many things easier and better.  That doesn't yet mean we
need to put all of them, or even some of them, in the tutorial.

> Regardless of the keyboard and input method issues, there are still
> three dominant encodings used in Germany and Austria, namely latin-1,
> windows-1252, and utf-8, and I consider it important that a user knows
> how to force an encoding in case Emacs fails to recognize it properly.

With the current Emacs, I don't think the user needs that knowledge as
badly as it was needed in v20.x or 21.x.  It might be that this
section was a must 10 or 8 years ago, but I think nowadays it just
gets in the way.

> As I said, the MULE section can be certainly improved :-)

I submit that MULE is not tutorial stuff.

So let's agree to disagree, and let Stefan and Chong make the decision
about this.

Re: TUTORIAL.de updated

[Ulrich Mueller]

>>>>> Werner LEMBERG wrote:

> I've done a quick update of TUTORIAL.de -- there are certainly
> issues which can be improved, but I'm short of time, unfortunately.

I've noticed that you've updated all `quotes' to »quotes« which is
certainly an improvement.

But why not „quotes“ which is the more common variant in German?

(BTW, are there any plans to stop that abuse of the grave accent as an
opening quotation mark in the rest of the documentation? Or am I the
only one who uses fonts where the `' pair looks so ugly and
asymmetric?)

Re: TUTORIAL.de updated

[Werner LEMBERG]

> I've noticed that you've updated all `quotes' to »quotes« which is
> certainly an improvement.
> 
> But why not „quotes“ which is the more common variant in German?

Because those quote characters are not available in Latin-1.
Additionally, I like » and « more :-) BTW, I doubt the `more common':
Virtually all novel books published in Germany or Austria use » and «.


    Werner

quotation marks [was Re: TUTORIAL.de updated]

[Glenn Morris]

Ulrich Mueller wrote:

> (BTW, are there any plans to stop that abuse of the grave accent as an
> opening quotation mark in the rest of the documentation? 

http://debbugs.gnu.org/cgi/bugreport.cgi?bug=10443

It will be a hard habit to change.

RE: quotation marks [was Re: TUTORIAL.de updated]

[Drew Adams]

> > (BTW, are there any plans to stop that abuse of the grave 
> > accent as an opening quotation mark in the rest of the
> > documentation?

I certainly hope not.  Quite misguided, IMO.

> http://debbugs.gnu.org/cgi/bugreport.cgi?bug=10443
> It will be a hard habit to change.

It is not just a question of habit.  Having different characters for starting
and ending such text simplifies distinguishing that text (e.g. font-locking).

Which start and end chars are conventionally chosen for this is not nearly as
important as the fact that they be different.

The URL that Glenn cites gives this as the (only) rationale for changing:

  This proposed patch follows up on recent changes to the
  GNU coding standards.  They now suggest that we should
  quote 'like this' or "like this" instead of `like this'; see
  <http://www.gnu.org/prep/standards/html_node/Quote-Characters.html>.

Such a change to "the GNU coding standards" might or might not make sense for C
coding (that is the only context cited where "the GNU coding standards" have
been changed in this regard).

But it definitely does not make sense for Emacs, IMO.  `...' is an Emacs doc &
UI convention, not simply a "coding standard".

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/19/12 09:22, Drew Adams wrote:
> `...' is an Emacs doc & UI convention, not simply a "coding standard".

For user interfaces it would be better if Emacs quoted
“like this” or ‘like this’ (i.e., using directed
quote marks) in the typical case where these quote marks
are supported.  Quoting `like this' is no longer appropriate,
since modern displays typically don't render ` and ' symmetrically,
and it confuses new users when Emacs gives unusual interpretations
to ` and '.

One way to address this would be to prefer directed
quote marks within the Emacs source code; the surrounding
technology should recognize these quotes and do the right
thing when displaying diagnostics on ASCII-only platforms.

Use of ` for Lisp backquotes, shell backquotes, etc., would
be undisturbed, of course.  The only change would be for
when we're quoting something in English text.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stephen J. Turnbull]

Paul Eggert writes:

 > Use of ` for Lisp backquotes, shell backquotes, etc., would
 > be undisturbed, of course.  The only change would be for
 > when we're quoting something in English text.

"Of course" is easy for a human to say; do you have code to implement
it automatically?

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/19/12 19:02, Stephen J. Turnbull wrote:
> "Of course" is easy for a human to say; do you have code to implement
> it automatically?

No, it will require a human-assisted pass over the sources, to
determine which instances of ` are for quoting English text,
and which are for other purposes.  This can be partly automated,
but in general some human judgment is required to change the
quoting style to one that's recommended by the GNU coding standards.

I've done this for other GNU projects; it wasn't that much work.
I expect Emacs will be somewhat harder, but it won't be too hard.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> Date: Thu, 19 Jan 2012 15:08:26 -0800
> From: Paul Eggert <eggert@cs.ucla.edu>
> Cc: 'Ulrich Mueller' <ulm@gentoo.org>, emacs-devel@gnu.org
> 
> On 01/19/12 09:22, Drew Adams wrote:
> > `...' is an Emacs doc & UI convention, not simply a "coding standard".
> 
> For user interfaces it would be better if Emacs quoted
> “like this” or ‘like this’ (i.e., using directed
> quote marks) in the typical case where these quote marks
> are supported.

Before we go that way, we will need to make sure the closing quote
mark behaves like ' does, wrt syntax (and thus word-level cursor
movement commands).  Currently it doesn't.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Alan Mackenzie]

Hello, Paul.

On Thu, Jan 19, 2012 at 03:08:26PM -0800, Paul Eggert wrote:
> On 01/19/12 09:22, Drew Adams wrote:
> > `...' is an Emacs doc & UI convention, not simply a "coding standard".

> For user interfaces it would be better if Emacs quoted
> ?like this? or ?like this? (i.e., using directed
> quote marks) in the typical case where these quote marks
> are supported.

Reading your mail across an SSH link with my ISP, I see these quote
marks as an inverted "?".  The same in mutt, where I am currently
typing, the same when I load it into Emacs.

> Quoting `like this' is no longer appropriate, since modern displays
> typically don't render ` and ' symmetrically, and it confuses new
> users when Emacs gives unusual interpretations to ` and '.

I think if ` and ' confuse new users, they perhaps should direct their
modest talents to a less demanding editor.  ;-)  Does anybody really
have trouble with a new style of quoting marks, assuming that they can
actually see them?  I never had trouble with, for example, « and »
(apologies if anybody can't see these.  They're double angle bracket
quote marks.)

I run Emacs on a Linux tty.  It's always been a requirement of Emacs
that it run OK in this environment.  

> One way to address this would be to prefer directed
> quote marks within the Emacs source code; the surrounding
> technology should recognize these quotes and do the right
> thing when displaying diagnostics on ASCII-only platforms.

What about displaying Elisp source code on ASCII-only platforms?

> Use of ` for Lisp backquotes, shell backquotes, etc., would
> be undisturbed, of course.  The only change would be for
> when we're quoting something in English text.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/19/12 23:49, Alan Mackenzie wrote:
> What about displaying Elisp source code on ASCII-only platforms?

This is not a new problem.  The Elisp source code in the Emacs trunk
currently has many non-ASCII characters.  Although most of these are
in the modules that deal with international characters, there are many
instances in otherwise-ordinary code, for example, the identifier
color-cie-ε in lisp/color.el, and the message containing the string
"the ‘+’ in this last example" in lisp/org/org-bibtex.el.

Despite the above, one can get lots of useful work done in an ASCII-only
environment when editing the Emacs source code.  This would continue
to be true even if Emacs used directed quotes in the source code more often.

> Reading your mail across an SSH link with my ISP, I see these quote
> marks as an inverted "?".  The same in mutt, where I am currently
> typing, the same when I load it into Emacs.

Yes, that's the default behavior if one is using the C locale.  These
days, though, it's typically nicer to use a locale like en_US.utf8
when developing Emacs code.  This works across an SSH link on a Linux
terminal -- I just now tried it, and the above non-ASCII examples worked
fine.  Although ASCII-only environments still exist, they're becoming
less important for Emacs development.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> Date: Fri, 20 Jan 2012 00:45:46 -0800
> From: Paul Eggert <eggert@cs.ucla.edu>
> Cc: 'Ulrich Mueller' <ulm@gentoo.org>, Drew Adams <drew.adams@oracle.com>,
> 	emacs-devel@gnu.org
> 
> Although ASCII-only environments still exist, they're becoming less
> important for Emacs development.

Not very long ago, some people here seemed to think otherwise:

  http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00445.html
  http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00447.html
  http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00451.html
  http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00518.html

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/20/2012 01:24 AM, Eli Zaretskii wrote:
>   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00445.html
>   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00447.html
>   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00451.html
>   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00518.html

Those messages are about using ASCII-only email, or having Emacs not
unexpectedly insert Unicode characters into an ASCII-only document,
or things of that sort, things that are about Emacs users in general.
Those are all valid issues, but they don't contradict the basic point
that it's becoming less important to cater to ASCII-only development
environments.

We routinely place more requirements on Emacs developers than on Emacs users,
and in particular we assume that Emacs developers can routinely view and
otherwise deal with non-ASCII characters when they need to.  This is not
an assumption we could have made 30 years ago, but it's a safe assumption
today.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Alan Mackenzie]

Hi, Paul.

On Fri, Jan 20, 2012 at 02:17:32PM -0800, Paul Eggert wrote:
> On 01/20/2012 01:24 AM, Eli Zaretskii wrote:

> Those messages are about using ASCII-only email, or having Emacs not
> unexpectedly insert Unicode characters into an ASCII-only document,
> or things of that sort, things that are about Emacs users in general.
> Those are all valid issues, but they don't contradict the basic point
> that it's becoming less important to cater to ASCII-only development
> environments.

Sorry, but that's a philosophical point that cannot pass without
challenge.  Because fewer people are using tty environments (which is
basically what is meant here by "ASCII-only") is it really "less
important" to support them?  Or is there some other reason?

There are at least three Emacs developers who use a tty rather than a
GUI.  There are users who do the same.  (Why use a GUI for pure text
work, indeed?)  To adapt a tty for random Unicode characters is a
non-trivial amount of work.

> We routinely place more requirements on Emacs developers than on Emacs users,
> and in particular we assume that Emacs developers can routinely view and
> otherwise deal with non-ASCII characters when they need to.  This is not
> an assumption we could have made 30 years ago, but it's a safe assumption
> today.

When I need to view one (which is rare), I use C-u C-x =.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> Date: Sat, 21 Jan 2012 08:59:05 +0000
> Cc: Eli Zaretskii <eliz@gnu.org>, ulm@gentoo.org, drew.adams@oracle.com,
>   emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> Sorry, but that's a philosophical point that cannot pass without
> challenge.  Because fewer people are using tty environments (which is
> basically what is meant here by "ASCII-only") is it really "less
> important" to support them?  Or is there some other reason?
> 
> There are at least three Emacs developers who use a tty rather than a
> GUI.  There are users who do the same.  (Why use a GUI for pure text
> work, indeed?)  To adapt a tty for random Unicode characters is a
> non-trivial amount of work.

TTY != pure-ASCII.

What kind of TTY do you use that needs a "non-trivial amount of work"
to support non-ASCII characters?  And what that is the work which
needs to be done for that?

Re: quotation marks [was Re: TUTORIAL.de updated]

[Alan Mackenzie]

On Sat, Jan 21, 2012 at 12:29:33PM +0200, Eli Zaretskii wrote:
> > Date: Sat, 21 Jan 2012 08:59:05 +0000
> > Cc: Eli Zaretskii <eliz@gnu.org>, ulm@gentoo.org, drew.adams@oracle.com,
> >   emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>

> > Sorry, but that's a philosophical point that cannot pass without
> > challenge.  Because fewer people are using tty environments (which is
> > basically what is meant here by "ASCII-only") is it really "less
> > important" to support them?  Or is there some other reason?

> > There are at least three Emacs developers who use a tty rather than a
> > GUI.  There are users who do the same.  (Why use a GUI for pure text
> > work, indeed?)  To adapt a tty for random Unicode characters is a
> > non-trivial amount of work.

> TTY != pure-ASCII.

Yes, OK.  But they share some of the problems WRT curly quote marks.

> What kind of TTY do you use that needs a "non-trivial amount of work"
> to support non-ASCII characters?  And what that is the work which
> needs to be done for that?

I have an ordinary Linux tty, currently running a Latin-1 character set.
I'd need somehow to arrange for the quotes to be added to that set.  I
don't know of any character set which includes both of these (I've had a
look through /usr/share/consolefonts/).  I don't know how to do this,
and finding out is going to be a lot of work.

Also, there's the issue of adapting the keyboard to support the new
characters.  I know how to do this, so it's no big deal for me.  But for
the vast majority of tty users, I'd guess this is another research
project.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> Date: Sat, 21 Jan 2012 11:12:38 +0000
> Cc: eggert@cs.ucla.edu, ulm@gentoo.org, drew.adams@oracle.com,
>   emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> I have an ordinary Linux tty, currently running a Latin-1 character set.

Is it possible to run it with the UTF-8 character set?  AFAIU, this
will solve all your problems, while still being compatible with Latin-1.

> Also, there's the issue of adapting the keyboard to support the new
> characters.

That's a different issue.  If and when we start using the non-ASCII
quotes, we will probably introduce some easy way of typing them
anyway.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Sat, 21 Jan 2012 11:12:38 +0000
>> Cc: eggert@cs.ucla.edu, ulm@gentoo.org, drew.adams@oracle.com,
>>   emacs-devel@gnu.org
>> From: Alan Mackenzie <acm@muc.de>
>> 
>> I have an ordinary Linux tty, currently running a Latin-1 character set.
>
> Is it possible to run it with the UTF-8 character set?  AFAIU, this
> will solve all your problems, while still being compatible with Latin-1.

A console font cannot contain more than 512 characters.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: quotation marks [was Re: TUTORIAL.de updated]

[David De La Harpe Golden]

On 21/01/12 12:52, Andreas Schwab wrote:
> Eli Zaretskii<eliz@gnu.org>  writes:
>
>>> Date: Sat, 21 Jan 2012 11:12:38 +0000
>>> Cc: eggert@cs.ucla.edu, ulm@gentoo.org, drew.adams@oracle.com,
>>>    emacs-devel@gnu.org
>>> From: Alan Mackenzie<acm@muc.de>
>>>
>>> I have an ordinary Linux tty, currently running a Latin-1 character set.
>>
>> Is it possible to run it with the UTF-8 character set?  AFAIU, this
>> will solve all your problems, while still being compatible with Latin-1.
>
> A console font cannot contain more than 512 characters.
>
> Andreas.

To expand a little, as, while that's true, that doesn't mean there's
no way to do it on an X-less system:

True text-mode (i.e. PC VGA gfx card) linux kernel VTs inherit that 
hardware limitation (256 glyphs with 16 colors or 512 glyphs with 8 
colors) [1] that the kernel can do relatively little about. You can work 
in unicode (see console-tools docs) ...but you can only set up a font 
showing a small subset of glyphs at a time.  The kernel VTs when on a 
framebuffer have some fairly similar limits IIRC, despite there not 
being a hardware reason in that case (lack of demand, compat, shared 
code, and keeping things simple spring to mind as possible non-hardware 
reasons, of course).

... However, there is nowadays a way around the limitation (presumably 
further reducing demand for any more advanced kernel support):

Use the userspace “fbterm” [2][3] terminal emulator instead of the linux 
kernel's more limited built-in. It is essentially a unicode-capable 
terminal emulator like those commonplace on X11 - but one that targets 
the framebuffer (and also now vesa modesetting, like (or perhaps using) 
old libsvga) rather than X11.

Though I currently use X11, so have little reason to use fbterm 
normally, I've just fairly successfully viewed the Hello file (and some 
curly quotes) in an emacs running under

“fbterm -n "DejaVu Sans Mono" -s 32 --vesa-mode=374“


[1] http://en.wikipedia.org/wiki/VGA_compatible_text_mode#Fonts
[2] http://code.google.com/p/fbterm/
[3] http://packages.debian.org/sid/fbterm

Re: quotation marks [was Re: TUTORIAL.de updated]

[David De La Harpe Golden]

On 22/01/12 03:27, David De La Harpe Golden wrote:
> “...“

hah. meant “...” of course.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Richard Stallman]

    There are at least three Emacs developers who use a tty rather than a
    GUI.  There are users who do the same.  (Why use a GUI for pure text
    work, indeed?)  To adapt a tty for random Unicode characters is a
    non-trivial amount of work.

I don't know how to do it.  I wish someone would fix these things
to "just work" in the future, but who knows how long that will take?

-- 
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
www.fsf.org  www.gnu.org
Skype: No way! That's nonfree (freedom-denying) software.
  Use free telephony http://directory.fsf.org/category/tel/

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/21/2012 12:59 AM, Alan Mackenzie wrote:

> Because fewer people are using tty environments (which is
> basically what is meant here by "ASCII-only") ...

No, by "ASCII-only" I meant displays that can display only ASCII.
I use tty environments all the time, and they're never ASCII-only.

> I have an ordinary Linux tty, currently running a Latin-1
> character set....  I don't know how to do this, and finding out
> is going to be a lot of work.

It shouldn't be a lot of work.  I don't normally use the Linux
console, but I looked into this problem and on my Ubuntu desktop
the fix is simple:

  export LC_ALL=en_US.utf8
  setupcon

After executing these two shell commands, I can see directed
quotes ‘like this’ with no problem, both within Emacs and when
catting to the terminal.  If it matters, I'm running Ubuntu 11.10
x86 on a 5-year-old desktop.

Does this approach work for you?

Re: quotation marks [was Re: TUTORIAL.de updated]

[Alan Mackenzie]

Morning, Paul.

On Sun, Jan 22, 2012 at 12:00:15AM -0800, Paul Eggert wrote:
> On 01/21/2012 12:59 AM, Alan Mackenzie wrote:

> > Because fewer people are using tty environments (which is
> > basically what is meant here by "ASCII-only") ...

> No, by "ASCII-only" I meant displays that can display only ASCII.
> I use tty environments all the time, and they're never ASCII-only.

OK.

> > I have an ordinary Linux tty, currently running a Latin-1
> > character set....  I don't know how to do this, and finding out
> > is going to be a lot of work.

> It shouldn't be a lot of work.  I don't normally use the Linux
> console, but I looked into this problem and on my Ubuntu desktop
> the fix is simple:

>   export LC_ALL=en_US.utf8
>   setupcon

> After executing these two shell commands, I can see directed
> quotes ‘like this’ with no problem, both within Emacs and when
> catting to the terminal.  If it matters, I'm running Ubuntu 11.10
> x86 on a 5-year-old desktop.

> Does this approach work for you?

So far I've been at it about 1½ hours and got nowhere.  I tried setting
LC_ALL (to en_UK.utf8), but haven't got `setupcon'.  So I tried Gentoo's
way, 

    # /etc/init.d/consolefont restart

, no joy.  Then I grepped through Gentoo's repository index for setupcon,
finding nothing.  Then wikipedia - Nothing.  Then Google - Hah!  I found
a man page.  This talks about /etc/default/console-setup, yet gives no
pointer to a description of that file.

At this point I get the familiar feeling of being on a wild goose chase,
which happens about every time I need to configure something new on Linux
(thankfully, GNU is much easier).

So a reversion to logic - whatever utility I use to set the font, that
font must exist in /usr/share/consolefonts.  But the font I'm already
using, lat1-16.psfu.gz, allegedly "[has] a built-in Unicode mapping",
whatever that might be.  Yet I don't get the curly quotes displayed in
Emacs (or anywhere else).

> "It shouldn't be a lot of work."

Maybe not, but there can be a large gap between "shouldn't be" and "won't
be".

Maybe you could give a run through of a hypothetical sequence of steps a
tty user would go through to find out how to set up his console for this.

Up to now it's taken nearly 2 hours now to get absolutely nowhere.  Maybe
I'm unusually inept at this sort of thing, but I don't think so.  I had a
head start by virtue of your help, for which thanks, but that misses the
point, that it will be very strenuous for a normal tty user to get set up
to see these curly quotes.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: quotation marks [was Re: TUTORIAL.de updated]

[James Cloos]

>>>>> "AM" == Alan Mackenzie <acm@muc.de> writes:

AM> So far I've been at it about 1½ hours and got nowhere.  I tried setting
AM> LC_ALL (to en_UK.utf8), but haven't got `setupcon'.  So I tried Gentoo's
AM> way, 

Terminus likely is the best font to use on the console for utf8:

    media-fonts/terminus-font with the default USE flags will work well.

In /etc/conf.d/consolefont set the consolefont variable.  I use:

    consolefont="ter-v20b"

the v stands for utf8, 20 is a 10x20 pixel font, b is bold (use n for
single pixel wide stems, b for two pixel wide stems).

I do not set consoletranslation or unicodemap in that file.

Set unicode="YES" in /etc/rc.conf.

The default /etc/env.d/02locale is:

LC_COLLATE="C"
LC_TIME="C"

You may want to add LANG=en_UK.UTF-8 to that.  Or add it to your $HOME
shell startup script.  I'd leave the two LC_ settings, though.  They
are important for portage.

You might want to set dumpkeys_charset in /etc/conf.d/keymaps.  The
comment explains it.

For that keymap config in that file, I find that emacs2 works quite well.
On gentoo that is stored in /usr/share/keymaps/i386/qwerty/emacs2.map.gz.
(The .map file is plain text.)  That said, you might prefer "uk".
Either way, though, I suspect that you'll find the quail input methods
to be the only good way to enter non-ascii on the console.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> Date: Fri, 20 Jan 2012 14:17:32 -0800
> From: Paul Eggert <eggert@cs.ucla.edu>
> CC: acm@muc.de, ulm@gentoo.org, drew.adams@oracle.com, 
>  emacs-devel@gnu.org
> 
> On 01/20/2012 01:24 AM, Eli Zaretskii wrote:
> >   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00445.html
> >   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00447.html
> >   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00451.html
> >   http://lists.gnu.org/archive/html/emacs-devel/2011-08/msg00518.html
> 
> Those messages are about using ASCII-only email, or having Emacs not
> unexpectedly insert Unicode characters into an ASCII-only document,
> or things of that sort, things that are about Emacs users in general.
> Those are all valid issues, but they don't contradict the basic point
> that it's becoming less important to cater to ASCII-only development
> environments.

Read the whole (longish) thread if you want to understand the context.
The context was precisely what is being discussed here: should we
seriously cater to ASCII-only environments, or can we forget about
them or treat them as second-class citizens.  The rest were examples
invoked to make this point or refute it.

> We routinely place more requirements on Emacs developers than on Emacs users,
> and in particular we assume that Emacs developers can routinely view and
> otherwise deal with non-ASCII characters when they need to.  This is not
> an assumption we could have made 30 years ago, but it's a safe assumption
> today.

I don't see how the difference between Emacs developers and users
comes into play here.  The change of the quotes is proposed for doc
strings, manuals, and other user-visible places.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Richard Stallman]

    > We routinely place more requirements on Emacs developers than on Emacs users,
    > and in particular we assume that Emacs developers can routinely view and
    > otherwise deal with non-ASCII characters when they need to.

I can "deal with" nonascii characters, but some of them look rather ugly.
Some of them cause the display to get out of whack, probably due to
mismatch between how wide they really display and how wide Emacs thinks
they are.

I don't want this to happen in my error messages and documentation!

Can you please send me a message containing the character proposed for this?

-- 
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
www.fsf.org  www.gnu.org
Skype: No way! That's nonfree (freedom-denying) software.
  Use free telephony http://directory.fsf.org/category/tel/

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/21/2012 10:30 PM, Richard Stallman wrote:

> I can "deal with" nonascii characters, but some of them look rather ugly.
> Some of them cause the display to get out of whack, probably due to
> mismatch between how wide they really display and how wide Emacs thinks
> they are.

That shouldn't happen with the quote characters we're talking about,
at least, it shouldn't happen on properly-configured platforms that
are commonly used today.

> Can you please send me a message containing the character proposed for this?

Sure:

‘foo’ -- directed single quotation marks (U+2018, U+2019)
“foo” -- directed double quotation marks (U+201C, U+201D)

If you're running a GNU/Linux distribution that works like Ubuntu,
and you're on the Linux console, the following shell commands
are meeded to configure the system properly for these characters.

  export LC_ALL=en_US.utf8
  setupcon

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stephen J. Turnbull]

Paul Eggert writes:

 >   setupcon

That would appear to be an Ubuntu-only command, or require a 3.x
kernel if it's a Linux utility.  I don't have it on either my Gentoo
system or my Debian(!) system, both of which are reasonably up-to-date
except for running late 2.6 kernels.

IOW, it's nto that easy, Paul.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Teemu Likonen]

* 2012-01-22T18:36:03+09:00 * Stephen J. Turnbull wrote:

> Paul Eggert writes:
>> setupcon
>
> That would appear to be an Ubuntu-only command, or require a 3.x
> kernel if it's a Linux utility. I don't have it on either my Gentoo
> system or my Debian(!) system, both of which are reasonably up-to-date
> except for running late 2.6 kernels.

In Debian the command /bin/setupcon is in the console-setup package.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Eli Zaretskii]

> From: "Stephen J. Turnbull" <stephen@xemacs.org>
> Date: Sun, 22 Jan 2012 18:36:03 +0900
> Cc: rms@gnu.org, ulm@gentoo.org, emacs-devel@gnu.org, acm@muc.de,
> 	Eli Zaretskii <eliz@gnu.org>, drew.adams@oracle.com
> 
> Paul Eggert writes:
> 
>  >   setupcon
> 
> That would appear to be an Ubuntu-only command, or require a 3.x
> kernel if it's a Linux utility.  I don't have it on either my Gentoo
> system or my Debian(!) system, both of which are reasonably up-to-date
> except for running late 2.6 kernels.

AFAICS, setupcon is a shell script that invokes either consolechars or
setfont (whichever it finds), and then invokes kbd_mode to set up the
keyboard input as well.  If you have those programs, you just need to
steal the script from somewhere.

(Caveat: I know absolutely nothing about this issue.)

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 1342 bytes --]

On 01/22/2012 01:36 AM, Stephen J. Turnbull wrote:
> I don't have it on either my Gentoo system or my Debian(!) system

It's from a Debian package, introduced in lenny (Debian GNU/Linux 5.0),
which was released about three years ago.  Here's a URL:

http://packages.debian.org/squeeze/console-setup

It's a GPLed shell script and is not that complicated -- it's a front
end to help out people who don't want to understand Linux consoles
(not that I blame them!).

If you don't have setupcon, you can try this recipe instead, as its
last two commands are what setupcon does, as far as fonts go:

  export LC_ALL=en_US.utf8
  printf '\033%%G'
  setfont -v /usr/share/consolefonts/Uni2-Fixed16.psf.gz

On 01/22/2012 02:22 AM, Alan Mackenzie wrote:
> Maybe you could give a run through of a hypothetical sequence of steps a
> tty user would go through to find out how to set up his console for this

Do you have setfont?  If so, do the above commands work?
I'll attach a copy of the psf.gz file.  (You can get its source from
the above URL.)

> it will be very strenuous for a normal tty user to get set up
> to see these curly quotes.

OK, but we're not talking about normal tty users.  We're talking about the
three Emacs developers who use Linux consoles to develop.  If we can
get a recipe working for them, then we've handled that problem.


[-- Attachment #2: Uni2-Fixed16.psf.gz --]
[-- Type: application/x-gzip, Size: 4112 bytes --]

Re: quotation marks [was Re: TUTORIAL.de updated]

[Alan Mackenzie]

Hello, Paul.

On Sun, Jan 22, 2012 at 02:54:47AM -0800, Paul Eggert wrote:
> On 01/22/2012 01:36 AM, Stephen J. Turnbull wrote:
> > I don't have it on either my Gentoo system or my Debian(!) system

> It's from a Debian package, introduced in lenny (Debian GNU/Linux 5.0),
> which was released about three years ago.  Here's a URL:

> http://packages.debian.org/squeeze/console-setup

> It's a GPLed shell script and is not that complicated -- it's a front
> end to help out people who don't want to understand Linux consoles
> (not that I blame them!).

> If you don't have setupcon, you can try this recipe instead, as its
> last two commands are what setupcon does, as far as fonts go:

>   export LC_ALL=en_US.utf8
>   printf '\033%%G'
>   setfont -v /usr/share/consolefonts/Uni2-Fixed16.psf.gz



> On 01/22/2012 02:22 AM, Alan Mackenzie wrote:
> > Maybe you could give a run through of a hypothetical sequence of steps a
> > tty user would go through to find out how to set up his console for this

> Do you have setfont?  If so, do the above commands work?
> I'll attach a copy of the psf.gz file.  (You can get its source from
> the above URL.)

Yes, the setfont command above did indeed work.  Presumably there will
be several such fonts corresponding with the different ascii/Latin-1
fonts.

For some reason, I didn't have that font in my /usr/share/consolefonts.

> > it will be very strenuous for a normal tty user to get set up
> > to see these curly quotes.

> OK, but we're not talking about normal tty users.  We're talking about the
> three Emacs developers who use Linux consoles to develop.  If we can
> get a recipe working for them, then we've handled that problem.

Are we?  I think I've lost track of this discussion.  I thought we were
talking about putting curly quotes into doc strings, etc, where ordinary
users are going to have to deal with them.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/22/2012 03:56 AM, Alan Mackenzie wrote:
> I thought we were talking about putting curly quotes into doc strings,
> etc, where ordinary users are going to have to deal with them.

No, the idea is doc strings would be automatically formatted
for the user's display, and that this would use ASCII quotes
unless Emacs is on a display that can handle curly quotes.
In other words, curly quotes would appear in the Emacs source,
but they're not necessarily what the Emacs user would see.

I am not proposing that .texi format be changed.  There's no need
for that, as Texinfo already generates different quoting styles
depending on the output technology, and its current input format
already suffices.

Nor am I worried about people ssh-ing from their cell phone to
develop Emacs.  If and when that becomes practical, we can fix it
for developers, much as we already have a fix for developers who
use Linux consoles.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stephen J. Turnbull]

Paul Eggert writes:
 > On 01/22/2012 01:36 AM, Stephen J. Turnbull wrote:
 > > I don't have it on either my Gentoo system or my Debian(!) system
 > 
 > It's from a Debian package, introduced in lenny (Debian GNU/Linux 5.0),
 > which was released about three years ago.  Here's a URL:
 > 
 > http://packages.debian.org/squeeze/console-setup

Well, the system was installed as stable in early 2007 and since
upgraded to testing (currently wheezy, plus some mission-critical
stuff from sid that I'm willing to baby-sit because it's mission
critical).  But that package is not installed on my system, which is a
common and up-to-date distribution.  It's possible that new installs
would always have it, but I suspect Emacs users are more likely to
have ancient repeatedly updated systems than the average user.

 > If you don't have setupcon, you can try this recipe instead,

It's not about me, OK?  If I want it, I doubt it would take me more
than half an hour to figure it out; I've been embedded in the world's
most complex linguistic environment (at least as far as computer use
goes) for 20 years, and was a recognized expert on the subject of
internationalization of GNU/Linux systems a decade ago, though I've
slipped quite a bit since then (obviously!)

I don't really have an objection to using directed quotes in Emacs, as
long as all the syntax tables and various things are set up correctly
before public release (and I would not be surprised to find out that
set up takes a while to shake down) -- at worst it should be annoying,
not work-stopping.  But I'd be willing to bet it's going to cause
annoyance to more than the three Emacs developers using the Linux
console if you start on it right now.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Lars Magne Ingebrigtsen]

"Stephen J. Turnbull" <stephen@xemacs.org> writes:

> But I'd be willing to bet it's going to cause annoyance to more than
> the three Emacs developers using the Linux console if you start on it
> right now.

If I ssh from my cell phone, and use Emacs under that, all unicode-ish
characters turn into four-letter monstrosities, which renders the text
unreadable.

I think it's about a decade too soon to think about making a change from
ASCII-mostly help texts in Emacs.  Maybe (just maybe) these things will
work better in 2022, and we can start discussing it again.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: quotation marks [was Re: TUTORIAL.de updated]

[Richard Stallman]

    ‘foo’ -- directed single quotation marks (U+2018, U+2019)
    “foo” -- directed double quotation marks (U+201C, U+201D)

On each line, both quotes display alike for me.
But at least they look like quotes.

setupcon gives me lots of errors; I guess it doesn't
work in this version.

-- 
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
www.fsf.org  www.gnu.org
Skype: No way! That's nonfree (freedom-denying) software.
  Use free telephony http://directory.fsf.org/category/tel/

Re: quotation marks

[Ulrich Mueller]

>>>>> On Fri, 20 Jan 2012, Paul Eggert wrote:

> The Elisp source code in the Emacs trunk currently has many
> non-ASCII characters. Although most of these are in the modules that
> deal with international characters, there are many instances in
> otherwise-ordinary code, for example, the identifier color-cie-ε in
> lisp/color.el, [...]

Hey, can we finally have “λ” instead of that clumsy “lambda” then? :-)

But seriously, I'd have expected that at least identifiers would stay
ASCII-only.

Re: quotation marks

[Juanma Barranquero]

On Fri, Jan 20, 2012 at 13:13, Ulrich Mueller <ulm@kph.uni-mainz.de> wrote:

> Hey, can we finally have “λ” instead of that clumsy “lambda” then? :-)

Thanks,  "(defalias 'λ 'lambda)" just entered my .emacs ;-)

    Juanma

RE: quotation marks

[Drew Adams]

> > Hey, can we finally have "?" instead of that clumsy 
> > "lambda" then? :-)
> 
> Thanks,  "(defalias '? 'lambda)" just entered my .emacs ;-)

Good idea.  And for display of existing `lambda' occurrencences:

http://www.emacswiki.org/emacs/PrettyLambda
http://www.emacswiki.org/emacs/pretty-lambdada.el 

Re: quotation marks

[Juanma Barranquero]

On Fri, Jan 20, 2012 at 16:51, Drew Adams <drew.adams@oracle.com> wrote:

> Good idea.  And for display of existing `lambda' occurrencences:

Cool, thanks.

    Juanma

Re: quotation marks [was Re: TUTORIAL.de updated]

[James Cloos]

Noting that ASCII ` always was a grave accent, and never a quote,
it seems that the proper fix for TeXinfo is to add one or more @foo
commands to cover quotations which are not already covered.

Adding @quote{} as a generic mechanism would be the starting point,
to specify locale-specific quoting.  Those locale details can be found
in the glibc src.  Separating semantics from presentation in document
sources is a beneficial goal.

Atoms with more specific semantics could be added in the furture.

(As a side note, Knuth(, et al?)’s choice of `' and ``'' instead of
a \command{} is one of the *very* few details I think he (they?) got
wrong.  I presume that TeXinfo’s usage derives from TeX’s.)

(It was most tempting to use Έλλενικ ΤεΧ instead of Latin TeX. :)

(I hope »Έλλενικ« is the correct declension.)

(And apologies for the abundance of parenthetical postscripts. ;)

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: quotation marks [was Re: TUTORIAL.de updated]

[Richard Stallman]

    Noting that ASCII ` always was a grave accent, and never a quote,
    it seems that the proper fix for TeXinfo is to add one or more @foo
    commands to cover quotations which are not already covered.

Texinfo uses ` and ' for quotes because that is the convention
in TeX.  This will not be changed.


-- 
Dr Richard Stallman
President, Free Software Foundation
51 Franklin St
Boston MA 02110
USA
www.fsf.org  www.gnu.org
Skype: No way! That's nonfree (freedom-denying) software.
  Use free telephony http://directory.fsf.org/category/tel/

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stefan Monnier]

Paul said:
> For user interfaces it would be better if Emacs quoted “like this” or
> ‘like this’ (i.e., using directed quote marks) in the typical case
> where these quote marks are supported.

I actually much prefer symmetry, the way it was with pre-unicode
X fonts.  E.g. ‛myfoo’.

> One way to address this would be to prefer directed
> quote marks within the Emacs source code; the surrounding
> technology should recognize these quotes and do the right
> thing when displaying diagnostics on ASCII-only platforms.

One other problem (besides display) with directed quote marks is that
they're not as easy/convenient to type.

Eli wrote:
> Before we go that way, we will need to make sure the closing quote
> mark behaves like ' does, wrt syntax (and thus word-level cursor
> movement commands).  Currently it doesn't.

While we'd probably want to fine tune this part, I don't think it's
a serious problem.  After all, the syntax of ' already varies depending
on the major mode, so it's already not always doing the right thing
(and I regularly bump into such problems where C-h f doesn't properly
get the function name from the text around point because of that), and
changing to directed quote marks could actually help with this problem.


        Stefan

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/20/2012 06:05 AM, Stefan Monnier wrote:
> I actually much prefer symmetry, the way it was with pre-unicode
> X fonts.  E.g. ‛myfoo’.
> ...
> One other problem (besides display) with directed quote marks is that
> they're not as easy/convenient to type.

Good points both.  Perhaps it would be nice to introduce a quoting
function that follows the user's preference without our having to
type directed quotes in the Elisp source code.  This could be a new
format flag `, e.g.,

  (format "Cannot open file %`s for reading", "foo.c")

could follow the user's preference for quoting, returning

  "Cannot open file ‛like this’ for reading"

in your preferred environment.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Ted Zlatanov]

On Fri, 20 Jan 2012 14:34:15 -0800 Paul Eggert <eggert@cs.ucla.edu> wrote: 

PE> Perhaps it would be nice to introduce a quoting function that
PE> follows the user's preference without our having to type directed
PE> quotes in the Elisp source code.  This could be a new format flag `,
PE> e.g.,

PE>   (format "Cannot open file %`s for reading", "foo.c")

PE> could follow the user's preference for quoting, returning

PE>   "Cannot open file ‛like this’ for reading"

PE> in your preferred environment.

That would be cool.  But maybe it would be nicer to have %` and %' for
open/close quote instead.

Ted

Re: quotation marks [was Re: TUTORIAL.de updated]

[Paul Eggert]

On 01/20/2012 05:13 PM, Ted Zlatanov wrote:
> maybe it would be nicer to have %` and %' for
> open/close quote instead

I had thought of that, but %' already has a different
meaning with POSIX printf (it inserts thousands grouping
characters when formatting numbers), and we might want to
implement that in Emacs someday.

We could use some other combinations, e.g., %< and %>
(they kinda look like French quote marks).  It's possible
to support both, that is, one could either write this:

  (format "Cannot open file %`s for reading" "foo.c")

or this:

  (format "Cannot open file %<%s%> for reading" "foo.c")

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stephen J. Turnbull]

Paul Eggert writes:
 > On 01/20/2012 05:13 PM, Ted Zlatanov wrote:
 > > maybe it would be nicer to have %` and %' for
 > > open/close quote instead

As Jim Cloos pointed out, we (at least he and I ;-) want to move in
the direction of semantic markup.  Also, I can just see people abusing
the one-sided quotes to get a "pretty" apostrophe.

 > I had thought of that, but %' already has a different
 > meaning with POSIX printf (it inserts thousands grouping
 > characters when formatting numbers), and we might want to
 > implement that in Emacs someday.
 > 
 > We could use some other combinations, e.g., %< and %>
 > (they kinda look like French quote marks).  It's possible
 > to support both,

You're getting carried away.  No thank you to multiple ways to do
something that's purely cosmetic.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Andreas Schwab]

"Stephen J. Turnbull" <stephen@xemacs.org> writes:

> Paul Eggert writes:
>  > We could use some other combinations, e.g., %< and %>
>  > (they kinda look like French quote marks).  It's possible
>  > to support both,
>
> You're getting carried away.  No thank you to multiple ways to do
> something that's purely cosmetic.

That's actually what GCC is using.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: quotation marks [was Re: TUTORIAL.de updated]

[Stephen J. Turnbull]

Andreas Schwab writes:
 > "Stephen J. Turnbull" <stephen@xemacs.org> writes:
 > 
 > > Paul Eggert writes:
 > >  > We could use some other combinations, e.g., %< and %>
 > >  > (they kinda look like French quote marks).  It's possible
 > >  > to support both,
 > >
 > > You're getting carried away.  No thank you to multiple ways to do
 > > something that's purely cosmetic.
 > 
 > That's actually what GCC is using.

GCC is using *both* %` to supply mirrored quotes *and* %<, %> to
supply individual quotes?  If so, -1 to GCC, too.

Re: quotation marks [was Re: TUTORIAL.de updated]

[Tom Tromey]

>>>>> "Stephen" == Stephen J Turnbull <stephen@xemacs.org> writes:

Stephen> GCC is using *both* %` to supply mirrored quotes *and* %<, %> to
Stephen> supply individual quotes?

Yes, you can use %< and %>, or use the 'q' flag to quote.

Stephen> If so, -1 to GCC, too.

It is convenient in practice.  %<..%> is more frequently used to quote
literal text, though I do see a few instances of %<%s%> in the tree,
where %qs would have been simpler.

Tom

Re: quotation marks [was Re: TUTORIAL.de updated]

[Ted Zlatanov]

On Sat, 21 Jan 2012 15:10:49 +0900 "Stephen J. Turnbull" <stephen@xemacs.org> wrote: 

SJT> Paul Eggert writes:
>> We could use some other combinations, e.g., %< and %>
>> (they kinda look like French quote marks).  It's possible
>> to support both,

SJT> You're getting carried away.  No thank you to multiple ways to do
SJT> something that's purely cosmetic.

It is as cosmetic as the thousands and decimal separators.  In a bg-UTF8
environment, for instance, the US defaults look very awkward.

Ted

Re: quotation marks [was Re: TUTORIAL.de updated]

[Ted Zlatanov]

On Fri, 20 Jan 2012 17:26:21 -0800 Paul Eggert <eggert@cs.ucla.edu> wrote: 

PE> On 01/20/2012 05:13 PM, Ted Zlatanov wrote:
>> maybe it would be nicer to have %` and %' for
>> open/close quote instead

PE> I had thought of that, but %' already has a different
PE> meaning with POSIX printf (it inserts thousands grouping
PE> characters when formatting numbers), and we might want to
PE> implement that in Emacs someday.

PE> We could use some other combinations, e.g., %< and %>
PE> (they kinda look like French quote marks).  It's possible
PE> to support both, that is, one could either write this:

PE>   (format "Cannot open file %`s for reading" "foo.c")

PE> or this:

PE>   (format "Cannot open file %<%s%> for reading" "foo.c")

That would be nice.

Ted

Re: TUTORIAL.de updated

[Stephen J. Turnbull]

Ulrich Mueller writes:
 > >>>>> Werner LEMBERG wrote:

 > (BTW, are there any plans to stop that abuse of the grave accent as an
 > opening quotation mark in the rest of the documentation? Or am I the
 > only one who uses fonts where the `' pair looks so ugly and
 > asymmetric?)

In printed documentation TeX will turn it into proper quotation
marks.  So, probably no change in Texinfo sources.

It should be possible to patch the to-Info and to-HTML drivers in
makeinfo to generate proper quotes in Latin-1.  Ask Karl Berry if a
patch would be accepted.

In docstrings, changing this would require teaching various parts of
the help system to recognize quotes rather than grave accent.  But
maybe help display functions (eg, C-h f) could do the translation
using display tables.

Re: quotation marks (was: Re: TUTORIAL.de updated)

[Ulrich Mueller]

>>>>> On Fri, 20 Jan 2012, Stephen J Turnbull wrote:

>> (BTW, are there any plans to stop that abuse of the grave accent as
>> an opening quotation mark in the rest of the documentation? Or am I
>> the only one who uses fonts where the `' pair looks so ugly and
>> asymmetric?)

> In printed documentation TeX will turn it into proper quotation
> marks.  So, probably no change in Texinfo sources.

Shouldn't it be written as @samp{foo} in Texinfo sources anyway?

> It should be possible to patch the to-Info and to-HTML drivers in
> makeinfo to generate proper quotes in Latin-1.  Ask Karl Berry if a
> patch would be accepted.

> In docstrings, changing this would require teaching various parts of
> the help system to recognize quotes rather than grave accent.  But
> maybe help display functions (eg, C-h f) could do the translation
> using display tables.

I think that display tables won't work, because some docstrings
contain code with (lisp) backquotes. I guess these shouldn't be shown
as quotes. ;-)

For example, this one:

   backquote is a Lisp macro in `backquote.el'.
   [...]
   `(a ,b c)      => (a (ba bb bc) c)	; insert the value of b
   `(a ,@b c)     => (a ba bb bc c)	; splice in the value of b

Re: quotation marks (was: Re: TUTORIAL.de updated)

[Eli Zaretskii]

> Date: Thu, 19 Jan 2012 19:50:03 +0100
> From: Ulrich Mueller <ulm@gentoo.org>
> Cc: emacs-devel@gnu.org
> 
> > In printed documentation TeX will turn it into proper quotation
> > marks.  So, probably no change in Texinfo sources.
> 
> Shouldn't it be written as @samp{foo} in Texinfo sources anyway?

I believe Stephen was talking about ``quoted text'', not about @samp
or @code or @file.