Number to Words and licencing

[Buddha Buck]

In the US at least it is customary to write out the amount and put it in
digits.  My checkbook in front of me has a form like:


Pay to the order of _________________________________________________ $
______
_______________________________________________________________ Dollars

I am expected to fill it out as:

Pay to the order of _John Smith_____________________________________
 $123.45
_One Hundred Twenty Three and 45
Cents----------------------------------------------__ Dollars

I would never expect that second line to be "One-Two-Three and Four-Five
Cents".  I believe I have seen some computer-generated checks which would
fill out the second line as "123.45****************" and not bother
converting it to spelled-out numbers.

I once made out a check for the amount "Twenty Even" and got charged $27,
so the amount in English does get read, even if incorrectly.  The rule is
that if the number (the $123.45) and the text (the "one hundred...")
disagree, the text takes precedent.




On Sun, Nov 3, 2013 at 9:25 PM, Johannes Kapune <listen at kapune.de> wrote:

> Am 04.11.2013 00:48, schrieb John Ralls:
>
>
>> On Nov 3, 2013, at 11:14 AM, Geert Janssens <janssens-geert at telenet.be>
>> wrote:
>>
>>  On Sunday 03 November 2013 08:01:38 John Ralls wrote:
>>>
>>>> On Nov 3, 2013, at 2:32 AM, Geert Janssens <janssens-geert at telenet.be>
>>>>
>>> wrote:
>>>
>>>> And that's only three languages that I know. I'm sure many other
>>>>> languages have still different ways to form larger numbers. These
>>>>> oddities in languages can't be covered in only with gettext as far
>>>>> as I know, or you will have to add quite a lot of individual
>>>>> numbers to translate.
>>>>>
>>>>
>>>> ICU has that built in and localized:
>>>> userguide.icu-project.org/formatparse/numbers
>>>> But note:
>>>> "ICU provides number spellout rules for several locales, but not for
>>>> all of the locales that ICU supports, and not all of the predefined
>>>> rule types. Also, as of release 2.6, some of the provided rules are
>>>> known to be incomplete."
>>>>
>>>> But I see no reason to add the dependency for something that's already
>>>> done via gettext.
>>>>
>>>>  This is the part I don't understand. How can gettext handle all the
>>> different number spellout rules ?
>>>
>>
>> Sorry, I misunderstood what you meant in " These oddities in languages
>> can't be covered in only with gettext as far as I know”, reading it as
>> “these oddities … can be covered only with gettext…”. Quite the opposite of
>> what you meant.
>>
>> I also failed to catch that in spite of all of the comments in
>> gnc-ui-utils.c, none of the strings are actually submitted to gettext,
>> which is explained with Christian’s comment about the integer_to_string
>> function being untranslatable.
>>
>> Sigh.
>>
>> So we *can* do that only in English.
>>
>> So I guess the question is do enough non-english speakers actually need
>> it? The Wikipedia article Thomas cited earlier suggests not, but perhaps
>> you have more direct knowledge otherwise.
>>
>> With that answered, do we want to bring in ICU as a dependency to meet
>> the requirement. The documentation of that class is not up to the standards
>> of some others, and I wasn’t able to figure out what locales are actually
>> supported.
>>
>>  Maybe I'm wrong but for printing on checks normally you print the digits
> as words: 123456,78 to
>
> ---one-two-three-four-five-six and seven-eight cents--- (per cent ...) or
> do you really write:
>
> onehundredtwentythreethousandfourhundredfiftysix Dollar and seventyeight
> cents?
>
> To make the checks save to manipulating the amount first sort of writing
> seems to me is enough and it is translatable to other languages.
>
> Best regards
> Johannes Kapune
>
>
> _______________________________________________
> gnucash-devel mailing list
> gnucash-devel at gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel
>