Here are some notes I think anyone who is programming in Japan may find useful, they (will) span across the full stack. At the time of posting this is a very small post, but this will be an ongoing project. I'll update it as I come back to issues again and again.
- Because Japanese has so many characters using hefty webfonts can significantly slow down your application. A good webfont, like Google Noto Sans CJK, has 65,535 glyphs! So,
- In general I recommend this old Stackoverflow post. But, here are two more links to useful articles on the topic: 1, 2.
- There are of course exceptions for stylistic design choices, and I've found Adobe Fonts are not so bad, they have an option for lazy/selective loading.
- For testing in a Ruby on Rails app: FFaker is better than Faker for dummy Japanese database input. This is an obscure reference, but it's just better. There's no additional setup, and to call a method with Japanese, it has it's own subclass, like: FFaker::AddressJA.method. It's a no brainer if you're using Rspec and FactoryBot.
- Β Namespaced i18n (example is a polymorphic namedspaced model)
# ja.yml
ja:
activerecord:
attributes:
namespace/address:
name: "εε"
company: "δΌη€Ύε"
address1: "ηΊεγ»ηͺε°"
address2: "ε»Ίη©εγ»ι¨ε±ηͺε·"
# some_view.html.erb
<%= form.fields_for :addresses, address do |address_fields| %>
<%= address_fields.label :name %>
<%= address_fields.text_field :name, class: 'form-control form-control-sm', value: address.name, placeholder: 'εε' %>