Re: [COMMITTED] fi_FI: Define yesstr, nostr

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Carlos O'Donell]

On 04/07/2013 10:18 AM, Keld Simonsen wrote:
> I think the yes and no strings should be starting with a lowercase letter.
> This in line with other names like day names and month names, they are also
> with initial lower case when that is the generic spelling.

You need to gather consensus on that and post a patch.

On 2013-03-21 Kalle Olavi Niemetalo <kon@iki.fi> wrote:
~~~
I had some doubts on whether the words should be capitalized.
Several other locales (including cs_CZ and tr_TR) use lower case
for yesstr and nostr, and "kyllä" and "ei" are normally written in
lower case in Finnish if they occur in the middle of a sentence.
However, I suppose yesstr and nostr are intended to be the entire
answer to some question asked by an application, rather than part
of a sentence, so "Kyllä" and "Ei" should be fine.
~~~

On 2013-03-21 Marko Myllyen <myllynen@redhat.com> wrote:
~~~
this is the rationale I used as well and the locales I happened to check
for reference (POSIX and en_IN) both had them capitalized. I would
suppose that if the words are in the middle of a sentence then they are
coming from an application which would have been translated themselves
and the plain yesstr/nostr are used in cases like you mentioned in your
follow-up email.
~~~

Consensus from Marko and Kalle seems to be that uppercase
letters should be the norm. I agree with that expectation for
at least Spanish (es_AR) and English (en_CA), two countries I
am intimately familiar with (being born in Argentina, but having
lived much of my life in Canada).

Can you make a case for not using uppercase?

Is it correct to compare fi_FI entries to other fi_FI entries?
* It seems to me that each entry is unique in the use it has for
programs.

Is it correct to compare fi_FI to other locales?
* This is probably the most rationale comparison since users
will be keeping the program the same, but switching locales.

Ask a native speaker?
* Perhaps the most sensible to ask the native speaker about
their expectations.

I see that en_US uses 'Yes' and 'No.'

That bolsters by expectation that all of the yesstr/nostr
values should be uppercase for use in a single word sentence
or cases (1), (2), and (3) as Kalle notes here:
http://sourceware.org/ml/libc-alpha/2013-03/msg00503.html

I'll be posting patches for es_AR and en_CA shortly which add
uppercase versions of yesstr/nostr.

Cheers,
Carlos.

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Carlos O'Donell]

On 04/07/2013 03:15 PM, Keld Simonsen wrote:
>> I don't know of any program actually using yesstr and nostr. 
> 
> Me neither. I was not aware that these keywords existed. Anyway, I
> will write some documentation for it.

I like your arguments. Could you help out the glibc community
by writing them into the wiki so we can use them as guidance
for future locale changes?

We have some initial information here:
http://sourceware.org/glibc/wiki/Locales

Cheers,
Carlos.

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Keld Simonsen]

On Sun, Apr 07, 2013 at 04:28:30PM -0400, Carlos O'Donell wrote:
> On 04/07/2013 03:15 PM, Keld Simonsen wrote:
> >> I don't know of any program actually using yesstr and nostr. 
> > 
> > Me neither. I was not aware that these keywords existed. Anyway, I
> > will write some documentation for it.
> 
> I like your arguments. Could you help out the glibc community
> by writing them into the wiki so we can use them as guidance
> for future locale changes?
> 
> We have some initial information here:
> http://sourceware.org/glibc/wiki/Locales

Yes, can do.

keld

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Keld Simonsen]

On Sun, Apr 07, 2013 at 11:35:35AM -0400, Carlos O'Donell wrote:
> On 04/07/2013 10:18 AM, Keld Simonsen wrote:
> > I think the yes and no strings should be starting with a lowercase letter.
> > This in line with other names like day names and month names, they are also
> > with initial lower case when that is the generic spelling.
> 
> You need to gather consensus on that and post a patch.
> 
> On 2013-03-21 Kalle Olavi Niemetalo <kon@iki.fi> wrote:
> ~~~
> I had some doubts on whether the words should be capitalized.
> Several other locales (including cs_CZ and tr_TR) use lower case
> for yesstr and nostr, and "kyllä" and "ei" are normally written in
> lower case in Finnish if they occur in the middle of a sentence.
> However, I suppose yesstr and nostr are intended to be the entire
> answer to some question asked by an application, rather than part
> of a sentence, so "Kyllä" and "Ei" should be fine.
> ~~~
> 
> On 2013-03-21 Marko Myllyen <myllynen@redhat.com> wrote:
> ~~~
> this is the rationale I used as well and the locales I happened to check
> for reference (POSIX and en_IN) both had them capitalized. I would
> suppose that if the words are in the middle of a sentence then they are
> coming from an application which would have been translated themselves
> and the plain yesstr/nostr are used in cases like you mentioned in your
> follow-up email.
> ~~~
> 
> Consensus from Marko and Kalle seems to be that uppercase
> letters should be the norm. I agree with that expectation for
> at least Spanish (es_AR) and English (en_CA), two countries I
> am intimately familiar with (being born in Argentina, but having
> lived much of my life in Canada).
> 
> Can you make a case for not using uppercase?

As said it is a design citeria for the locales, both
POSIX and TR 14652 and TR 30112 that if a word in a language canonically begins
with a small letter, then the locale has the same word with an initial
small letter. This is accordance with normal dictionalry stylei, eg. OED.
And also with definition style in ISO standards. Then it is up to
the application to capitalize the initial letter if this is required
eg because it is the first word in a sentence.

Eg. the English OK is written in Danish as "o.k." according
to Danish ortography. Specifying it as "O.k." would not reflect 
this rule.

> 
> Is it correct to compare fi_FI entries to other fi_FI entries?
> * It seems to me that each entry is unique in the use it has for
> programs.
> 
> Is it correct to compare fi_FI to other locales?
> * This is probably the most rationale comparison since users
> will be keeping the program the same, but switching locales.
> 
> Ask a native speaker?
> * Perhaps the most sensible to ask the native speaker about
> their expectations.
> 
> I see that en_US uses 'Yes' and 'No.'
> 
> That bolsters by expectation that all of the yesstr/nostr
> values should be uppercase for use in a single word sentence
> or cases (1), (2), and (3) as Kalle notes here:
> http://sourceware.org/ml/libc-alpha/2013-03/msg00503.html

That posting included:

> I meant yesstr and nostr could be used in these ways:
> 
> (1) A program writes a question to stdout, disables tty echo, and
>    waits for an answer.  When the user presses "k", the program
>    takes that as "Yes" and echoes the whole string "Kyllä".  A
>    capital K is fine here.

I think it is even or more natural to echo back "k", as this is most likely
what the user typed in. Users normally type in lowercase.


> (2) A program writes a question to stdout and appends "(Kyllä/Ei)".
>    Upper-case letters are good here because they match the
>    characters that yesexpr and noexpr recognize.

Normal practice in linux interfaces is to write with case the possible
answers. If you write "Kyllä/Ei" then people would expect to answer with a capital K or capital E.
And that would be cumbersome - needing two fingers in operation in stead of just one.

> (3) A program displays "Kyllä" and "Ei" buttons in a window and
>    expects the user to click one of them.  Such buttons
>    typically start with upper-case letters.

That is probably true. Here you should then convert the first letter to uppercase.


> (4) A program displays a list of key-value pairs, something like
>      Ilmastointi: kyllä
>      CD-soitin:   kyllä
>      Vetokoukku:  ei
>      Väri:        punainen
>    Here, lower case might be more common in Finnish.

Yes, and probably also in Danish and English.

> I don't know of any program actually using yesstr and nostr. 

Me neither. I was not aware that these keywords existed. Anyway, I will write some documentation for it.

Best regards
Keld

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Andreas Jaeger]

On Sunday, April 07, 2013 21:15:56 Keld Simonsen wrote:
> On Sun, Apr 07, 2013 at 11:35:35AM -0400, Carlos O'Donell wrote:
> [...]
> > I don't know of any program actually using yesstr and nostr.
> 
> Me neither. I was not aware that these keywords existed. Anyway, I
> will write some documentation for it.

We have some usage in nscd/nscd_stat.c of yesstr and nostr ;)

Andreas
-- 
 Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
  SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
   GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
    GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Troy Korjuslommi]

Requiring a programmer to convert initial characters to upper/lower case
moots the point of having locale data. Many languages (e.g. Japanese,
Chinese) don't even have upper and lower case. How is a programmer
supposed to know generalities about all the world's languages? Isn't
that why we have locale data?

When writing an app, I have no idea how some obscure language (e.g.
Finnish, which is not obscure to me, but is to most programmers) uses
upper and lower case letters.

I think the key point is to document the expected usage of key values
properly. When use cases are specified explicitly, translators can write
good values for key values, and programmers know when using a value is
appropriate and when not. If a programmer encounters a situation not
documented anywhere, he/she can post a request to add such a key value
to CLDR and to glibc.

Also, just because we don't have referable programs which use yesstr
etc., doesn't mean many people don't use them, and expect them to be
correct for their programs.

As an aside, if Marko says capitalization is appropriate in this case, I
trust his judgement. He is a member of the Kotoistus team and is able to
ask Finnish language experts for confirmation for his case, so at least
from a usage, linguistic and grammatical perspective he has certain
authority in the matter.


Troy Korjuslommi





On Sun, 2013-04-07 at 21:15 +0200, Keld Simonsen wrote:
> On Sun, Apr 07, 2013 at 11:35:35AM -0400, Carlos O'Donell wrote:
> > On 04/07/2013 10:18 AM, Keld Simonsen wrote:
> > > I think the yes and no strings should be starting with a lowercase letter.
> > > This in line with other names like day names and month names, they are also
> > > with initial lower case when that is the generic spelling.
> > 
> > You need to gather consensus on that and post a patch.
> > 
> > On 2013-03-21 Kalle Olavi Niemetalo <kon@iki.fi> wrote:
> > ~~~
> > I had some doubts on whether the words should be capitalized.
> > Several other locales (including cs_CZ and tr_TR) use lower case
> > for yesstr and nostr, and "kyllä" and "ei" are normally written in
> > lower case in Finnish if they occur in the middle of a sentence.
> > However, I suppose yesstr and nostr are intended to be the entire
> > answer to some question asked by an application, rather than part
> > of a sentence, so "Kyllä" and "Ei" should be fine.
> > ~~~
> > 
> > On 2013-03-21 Marko Myllyen <myllynen@redhat.com> wrote:
> > ~~~
> > this is the rationale I used as well and the locales I happened to check
> > for reference (POSIX and en_IN) both had them capitalized. I would
> > suppose that if the words are in the middle of a sentence then they are
> > coming from an application which would have been translated themselves
> > and the plain yesstr/nostr are used in cases like you mentioned in your
> > follow-up email.
> > ~~~
> > 
> > Consensus from Marko and Kalle seems to be that uppercase
> > letters should be the norm. I agree with that expectation for
> > at least Spanish (es_AR) and English (en_CA), two countries I
> > am intimately familiar with (being born in Argentina, but having
> > lived much of my life in Canada).
> > 
> > Can you make a case for not using uppercase?
> 
> As said it is a design citeria for the locales, both
> POSIX and TR 14652 and TR 30112 that if a word in a language canonically begins
> with a small letter, then the locale has the same word with an initial
> small letter. This is accordance with normal dictionalry stylei, eg. OED.
> And also with definition style in ISO standards. Then it is up to
> the application to capitalize the initial letter if this is required
> eg because it is the first word in a sentence.
> 
> Eg. the English OK is written in Danish as "o.k." according
> to Danish ortography. Specifying it as "O.k." would not reflect 
> this rule.
> 
> > 
> > Is it correct to compare fi_FI entries to other fi_FI entries?
> > * It seems to me that each entry is unique in the use it has for
> > programs.
> > 
> > Is it correct to compare fi_FI to other locales?
> > * This is probably the most rationale comparison since users
> > will be keeping the program the same, but switching locales.
> > 
> > Ask a native speaker?
> > * Perhaps the most sensible to ask the native speaker about
> > their expectations.
> > 
> > I see that en_US uses 'Yes' and 'No.'
> > 
> > That bolsters by expectation that all of the yesstr/nostr
> > values should be uppercase for use in a single word sentence
> > or cases (1), (2), and (3) as Kalle notes here:
> > http://sourceware.org/ml/libc-alpha/2013-03/msg00503.html
> 
> That posting included:
> 
> > I meant yesstr and nostr could be used in these ways:
> > 
> > (1) A program writes a question to stdout, disables tty echo, and
> >    waits for an answer.  When the user presses "k", the program
> >    takes that as "Yes" and echoes the whole string "Kyllä".  A
> >    capital K is fine here.
> 
> I think it is even or more natural to echo back "k", as this is most likely
> what the user typed in. Users normally type in lowercase.
> 
> 
> > (2) A program writes a question to stdout and appends "(Kyllä/Ei)".
> >    Upper-case letters are good here because they match the
> >    characters that yesexpr and noexpr recognize.
> 
> Normal practice in linux interfaces is to write with case the possible
> answers. If you write "Kyllä/Ei" then people would expect to answer with a capital K or capital E.
> And that would be cumbersome - needing two fingers in operation in stead of just one.
> 
> > (3) A program displays "Kyllä" and "Ei" buttons in a window and
> >    expects the user to click one of them.  Such buttons
> >    typically start with upper-case letters.
> 
> That is probably true. Here you should then convert the first letter to uppercase.
> 
> 
> > (4) A program displays a list of key-value pairs, something like
> >      Ilmastointi: kyllä
> >      CD-soitin:   kyllä
> >      Vetokoukku:  ei
> >      Väri:        punainen
> >    Here, lower case might be more common in Finnish.
> 
> Yes, and probably also in Danish and English.
> 
> > I don't know of any program actually using yesstr and nostr. 
> 
> Me neither. I was not aware that these keywords existed. Anyway, I will write some documentation for it.
> 
> Best regards
> Keld

Re: [COMMITTED] fi_FI: Define yesstr, nostr

[Keld Simonsen]

I agree that programmers should not need to know the specifics of the
different languages.

However, I think (but I am not sure...) that
all languages that have the uppercase and lowercase distinction, use uppercase in the
beginning of a sentence, and this is one of the main purposes for the
yesstr/nostr to be uppercase or lowercase initially. There are
functions and even format specifiers in a number of languages
to specify that the first letter needs to be converted to uppercase when it is supposed
to be the beginning word of a sentence. I believe we only have
uppercase/lowercase for the scripts Latin, Greek and Cyrillic. so it would be safe
for a programmer to specify this with a format specifier.

Anyway Marko said that in some cases Finnish would use uppercase for this,
and in some cases not. 

And what I wrote was that there are design policies for glibc and locales that we
should try to follow.  And behind these philosophies there are traditions 
that are followed in language ortographies, which are common amongst the 
languages affected.

best regards
keld


On Tue, Apr 09, 2013 at 10:43:08AM +0300, Troy Korjuslommi wrote:
> Requiring a programmer to convert initial characters to upper/lower case
> moots the point of having locale data. Many languages (e.g. Japanese,
> Chinese) don't even have upper and lower case. How is a programmer
> supposed to know generalities about all the world's languages? Isn't
> that why we have locale data?
> 
> When writing an app, I have no idea how some obscure language (e.g.
> Finnish, which is not obscure to me, but is to most programmers) uses
> upper and lower case letters.
> 
> I think the key point is to document the expected usage of key values
> properly. When use cases are specified explicitly, translators can write
> good values for key values, and programmers know when using a value is
> appropriate and when not. If a programmer encounters a situation not
> documented anywhere, he/she can post a request to add such a key value
> to CLDR and to glibc.
> 
> Also, just because we don't have referable programs which use yesstr
> etc., doesn't mean many people don't use them, and expect them to be
> correct for their programs.
> 
> As an aside, if Marko says capitalization is appropriate in this case, I
> trust his judgement. He is a member of the Kotoistus team and is able to
> ask Finnish language experts for confirmation for his case, so at least
> from a usage, linguistic and grammatical perspective he has certain
> authority in the matter.
> 
> 
> Troy Korjuslommi
> 
> 
> 
> 
> 
> On Sun, 2013-04-07 at 21:15 +0200, Keld Simonsen wrote:
> > On Sun, Apr 07, 2013 at 11:35:35AM -0400, Carlos O'Donell wrote:
> > > On 04/07/2013 10:18 AM, Keld Simonsen wrote:
> > > > I think the yes and no strings should be starting with a lowercase letter.
> > > > This in line with other names like day names and month names, they are also
> > > > with initial lower case when that is the generic spelling.
> > > 
> > > You need to gather consensus on that and post a patch.
> > > 
> > > On 2013-03-21 Kalle Olavi Niemetalo <kon@iki.fi> wrote:
> > > ~~~
> > > I had some doubts on whether the words should be capitalized.
> > > Several other locales (including cs_CZ and tr_TR) use lower case
> > > for yesstr and nostr, and "kyllä" and "ei" are normally written in
> > > lower case in Finnish if they occur in the middle of a sentence.
> > > However, I suppose yesstr and nostr are intended to be the entire
> > > answer to some question asked by an application, rather than part
> > > of a sentence, so "Kyllä" and "Ei" should be fine.
> > > ~~~
> > > 
> > > On 2013-03-21 Marko Myllyen <myllynen@redhat.com> wrote:
> > > ~~~
> > > this is the rationale I used as well and the locales I happened to check
> > > for reference (POSIX and en_IN) both had them capitalized. I would
> > > suppose that if the words are in the middle of a sentence then they are
> > > coming from an application which would have been translated themselves
> > > and the plain yesstr/nostr are used in cases like you mentioned in your
> > > follow-up email.
> > > ~~~
> > > 
> > > Consensus from Marko and Kalle seems to be that uppercase
> > > letters should be the norm. I agree with that expectation for
> > > at least Spanish (es_AR) and English (en_CA), two countries I
> > > am intimately familiar with (being born in Argentina, but having
> > > lived much of my life in Canada).
> > > 
> > > Can you make a case for not using uppercase?
> > 
> > As said it is a design citeria for the locales, both
> > POSIX and TR 14652 and TR 30112 that if a word in a language canonically begins
> > with a small letter, then the locale has the same word with an initial
> > small letter. This is accordance with normal dictionalry stylei, eg. OED.
> > And also with definition style in ISO standards. Then it is up to
> > the application to capitalize the initial letter if this is required
> > eg because it is the first word in a sentence.
> > 
> > Eg. the English OK is written in Danish as "o.k." according
> > to Danish ortography. Specifying it as "O.k." would not reflect 
> > this rule.
> > 
> > > 
> > > Is it correct to compare fi_FI entries to other fi_FI entries?
> > > * It seems to me that each entry is unique in the use it has for
> > > programs.
> > > 
> > > Is it correct to compare fi_FI to other locales?
> > > * This is probably the most rationale comparison since users
> > > will be keeping the program the same, but switching locales.
> > > 
> > > Ask a native speaker?
> > > * Perhaps the most sensible to ask the native speaker about
> > > their expectations.
> > > 
> > > I see that en_US uses 'Yes' and 'No.'
> > > 
> > > That bolsters by expectation that all of the yesstr/nostr
> > > values should be uppercase for use in a single word sentence
> > > or cases (1), (2), and (3) as Kalle notes here:
> > > http://sourceware.org/ml/libc-alpha/2013-03/msg00503.html
> > 
> > That posting included:
> > 
> > > I meant yesstr and nostr could be used in these ways:
> > > 
> > > (1) A program writes a question to stdout, disables tty echo, and
> > >    waits for an answer.  When the user presses "k", the program
> > >    takes that as "Yes" and echoes the whole string "Kyllä".  A
> > >    capital K is fine here.
> > 
> > I think it is even or more natural to echo back "k", as this is most likely
> > what the user typed in. Users normally type in lowercase.
> > 
> > 
> > > (2) A program writes a question to stdout and appends "(Kyllä/Ei)".
> > >    Upper-case letters are good here because they match the
> > >    characters that yesexpr and noexpr recognize.
> > 
> > Normal practice in linux interfaces is to write with case the possible
> > answers. If you write "Kyllä/Ei" then people would expect to answer with a capital K or capital E.
> > And that would be cumbersome - needing two fingers in operation in stead of just one.
> > 
> > > (3) A program displays "Kyllä" and "Ei" buttons in a window and
> > >    expects the user to click one of them.  Such buttons
> > >    typically start with upper-case letters.
> > 
> > That is probably true. Here you should then convert the first letter to uppercase.
> > 
> > 
> > > (4) A program displays a list of key-value pairs, something like
> > >      Ilmastointi: kyllä
> > >      CD-soitin:   kyllä
> > >      Vetokoukku:  ei
> > >      Väri:        punainen
> > >    Here, lower case might be more common in Finnish.
> > 
> > Yes, and probably also in Danish and English.
> > 
> > > I don't know of any program actually using yesstr and nostr. 
> > 
> > Me neither. I was not aware that these keywords existed. Anyway, I will write some documentation for it.
> > 
> > Best regards
> > Keld
>