Validation and normalization of Zenkaku Phone Numbers

Japanese and some other Asian scripts have full and half-width numbers. We should normalize and validate these accordingly.

For example, from the NTT web pages, we have a company with a phone number 0800-8000800 (hankaku) but when we go to the company home page, it's written as (zenkaku). We should probably allow both and either normalize zenkaku to hankaku, or tweak the validator to allow both. FWIW. Most of the real validation should probably be done in the browser, where the locale is known.

A phone number might also be written using kanji digits, but that would be unusual.

FWIW. Jira throws an exception when I enter the zenkaku phone number in this field...