Less free underscores in number literals

[KennyTM~]

On Oct 24, 10 03:11, Rainer Deyke wrote:
> On 10/23/2010 10:53, Kagamin wrote:
>> It's how their language builds numbers. Numbers written in ideographs
>> use this grouping, but this doesn't mean, they use the same grouping
>> for arabic digits. For example, amazon.co.jp uses arabic numbers and
>> western 3-digit grouping.
>
> Using groupings of three digits in Japanese seems extremely awkward,
> especially for larger numbers, since you would have to mentally regroup
> the digits in groups of four in order to read it.  It's not just the
> written language but the spoken language that uses groups of four.  For
> example, the number 1,234,567,890 would be read as 12億, 3456万, 7890.
>
> If amazon.jp uses groups of three, then my initial reaction is
> "imperfect localization".
>

[Off topic]

While it is read in groups of 4, I've never seen the number grouped by 4 
digits in any part of East Asia. It is enter written as groups of three 
"1,234,567,890", without grouping "1234567890" or using the native units 
like "12億", "十二億三千四百五十六万七千八百九十" etc.

> Also, even in English there are cases where groupings other than three
> make sense.  Consider:
>
> int price_in_cents = 54_95;
>
>