Home / ubuntu / Questions / 1144691
In Process
Nicolas Raoul
Asked: 2019-05-21 00:54:14 +0800 CST

How to set up Ubuntu to less easily make full-width/half-width number mistakes?

  • 772

In Japanese, numbers can be written either full-width or half-width:

  • Full-width: １２３４５６７８９０
  • Half-width: 1234567890

Nowadays in my field everyone use half-width, and using full-width is seen as a gross lack of professionalism. Unfortunately Ubuntu makes it easy to mix up full-width and half-width.

Some programs make the distinction very clear, for instance vi: enter image description here

... but other programs (most importantly firefox) make almost no distinction:

enter image description here

That leads to very incompetent-looking messages. When the customers look at the message above in their Mac browsers the distinction is super-clear.

Question

How to set up Ubuntu to make it much more difficult to mistakenly type full-width numbers?

I am not sure what is the root problem, but I can imagine two strategies:

  • Prevent programs from using fonts that do not make the distinction clear enough.
  • Modify the input method to never input full-width numbers.

I have no idea how to implement these strategies though. Other strategies are also very welcome!

firefox fonts japanese mozc

1 Answers

  1. The full-width characters are typed only when you attempt to input 2-byte characters, which are mostly Asian languages, I think.

    This is an example how you avoid to type in a full-width characters, especially numbers.

    enter image description here

    This is a mozc configuration panel; advanced tab from which is brought by this way.

    enter image description here

    There're many other input methods and should be set by different ways but almost always the same, e.g. Anthy or ATOK, Windows IME and etc.

    The other way to force full-width numbers replaced into half-width, when your text file probably contains some mixed full- and half-width numbers.

    sed 'y/０１２３４５６７８９/0123456789/'

    For example, you have an uncertain text1 file looks like:
    12123 １５３２ ４５６

    And you want to replace all the full-width numbers into half-width numbers and create a new file text2

    cat text1 | sed 'y/０１２３４５６７８９/0123456789/' > text2

    You get:
    12123 1532 456

    ps. You can also replace full-width alphabet letters into half-width, that would be:

    sed 'y/０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ/0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz/'
    • 0