Keyboard layouts and CLDR

[Marcel Schneider via Unicode]

On Tue, 30 Jan 2018 08:18:49 +0100, Philippe Verdy wrote:
> 
> I have always wondered why Microsoft did not push itself at least the five
> simple additions needed since long in French for the French AZERTY LAYOUT:

Many people in Fɽanƈë are wondering, but it is primarily a matter of honoring 
a countryʼs policies and not interfering with official work. France is expected to 
fix itself its keyboarding problems and publish a standard, and thatʼs what is 
actually happening. See Shawn Steeleʼs blog post about Locale Data in Windows 10 & CLDR:

https://blogs.msdn.microsoft.com/shawnste/2015/08/29/locale-data-in-windows-10-cldr/

> - [AltGr]+[²] to produce the cedilla dead key (needed only before capital C in French) :
> this is frequently needed, the alternative would be [AltGr]+[C] to map "Ç" without the dead key;

That would be easier than Alt+0199. But Alt+something should yield consistently either uppercase 
or lowercase. Then, especially in the United States, not having all uppercase letters accessed with 
Shift+lowercase is considered counter-intuitive. And when the lowercase letter is directly accessed, 
going through a dead key to get its uppercase is not something I would recommend. Therefore, 
all uppercase that are used as initials (not Ù) should be Shift+lowercase, and digits in AltGr like on 
a few Latin layouts shipping with Windows, plus a Programmer toggle described in:

https://unicode.org/cldr/trac/ticket/10851#comment:2

> spell checkers forget the frequent words: Ça or Ç'.

I never use spell checkers, and when they show up with red wavy underline, I quickly try to disable 
them (outside of Gooogle Search). That is why I have typos. [This one has occurred unintentionally.]

>
> - [AltGr]+[1&] to produce the acute accent dead key (similar to [AltGr+7è`] giving the grave accent deadkey) :
> this is the most frequent missing letter we need all the time.

Therefore, the É should be mapped to a live key. But the acute dead key is really the missing one. 
Belgiumʼs AZERTY has it. Getting É at least by dead key would have divided our trouble by half.

>
> - [AltGr]+[O] to produce "œ" (without ShiftLock or CapsLock mode enabled),
> or "Œ" (in ShiftLock or CapsLock mode), and
> [AltGr]+[Shift]+[O] to produce "Œ" (independantly of [ShiftLock] which is disabled by [Shift], but without [CapsLock])
> or "œ" (independantly of [CapsLock], but without [ShiftLock]) :
> this is needed occasionnaly for very few common words, the most frequent omission is "Œuf" or its plural "Œufs".

To repay the Œœ for its exclusion from Latin-1 (due to a Frenchman), it should be granted 
two key positions in the Base and Shift shift states, amidst the upper row letters.

>
> - [AltGr]+[A] to produce "æ" (without ShiftLock or CapsLock mode enabled),
> or "Æ" (in ShiftLock or CapsLock mode), and
> [AltGr]+[Shift]+[O] to produce "Æ" (independantly of [ShiftLock] which is disabled by [Shift], but without [CapsLock])
> or "æ" (independantly of [CapsLock], but without [ShiftLock]) :
> this is rarely needed, except for a few words borrowed from Latin used in biology or some legal/judiciary terminology.

And one spelling of _Lætitia_.

>
> - Adding Y to the list of allowed letters after the dieresis deadkey to produce "Ÿ" :
> the most frequent case is L'HAŸE-LÈS-ROSES, the official name of a French municipality when written with full capitalisation,
> almost all spell checkers often forget to correct capitalized names such as this one.

Thatʼs really something I never understood neither. Why that deadlist was not updated. 
Maybe like above: If Microsoft had updated our layout with 'Ÿ', we could have wondered 
why they didnʼt add the other missing stuff while they were on it.

>
> This would allow typing French completely including on initial capitals.
> All other French capital letters can be typed (ÂÊÎÔÛ with the circumflex dead key,
> ËÏÜŸ with the dieresis dead key which already allows ÄÖ not needed for French but for Alsatian or some names borrowed from German).
>
> But we have mappings already in the AZERTY layout for:
> - the tilde as a dead key on [AltGr]+[2é~], even if it is not used for French but only for "ñ" or "Ñ" in names from Spanish or Breton,

That didnʼt prevent Breton authorities from refusing it in a first name, 
Denis Jacquerye reported in the wake of the Kazakh apostrophe thread:

http://unicode.org/mail-arch/unicode-ml/y2018-m01/0133.html

> " ÃÕ" not needed at all, /ãõ/ needed only for standard French IPA phonetics where we still can't type /ɑɡʀɔɲ/ for French phonetics
> - the grave accent as a dead key on [AltGr]+[7è`], needed for "ÀÈ" but allowing also "ÌÒÙ" not used at all in French.
>
> There's not any good rationale in the French AZERTY layout to keep it incomplete on capitals
> while maintaining other capital letters with diacritics composed with dead keys but not needed at all in French,
> except the case of "ŸœŒ" missing from ISO 8859-1 but present in Windows-1252.

There is even a way of putting all into the existing dead keys, if ‹ circumflex accent › (that is our directly accessed 
dead key) followed by any diacriticized letter did yield its uppercase, and followed by b or q, yield æ or œ… 
But that isnʼt what one would call a properly designed keyboard layout.

> 
----
> Using the Windows "Charmap" accessory with the "Unicode" charset and "Latin" subset is still too difficult to locate the missing letters,
> as it is only sorted by code point value but still does not cover all Latin letters;
> the Windows "Charmap" tool is usable for French only when selecting the Windows-1252 charset (aka "Windows : Occidental").
>
> But I don't understand why this accessory cannot simply add some rows at top of the table for the current language selected
> on the "Languages Bar", or why it does not simply features the complete alphabet of the current language, sorted correctly
> according to CLDR rules for that language (not sorted randomly by code point value) to make it really usable. If we select
> another subset, it should also be sortable according to language rules (or CLDR default root otherwise) and not according to code point value: 
> this could be a simple checkbox or a pair of radio buttons (binary sort, or alphabetic sort). 
>
> Finally, the Charmap tool should be updated to add missing characters that are not covered in the "Unicode" charset selection,
> even if they are encoded in Unicode and really mapped in fonts: the coverage of proposed "subsets" is an extremely old version of Unicode.

I see that as a very valuable feature request. And this one doesnʼt need to wait for any national standard 
to get implemented.

Regards,

Marcel