Please avoid creation of tags with full-width alphanumeric characters

Question

At Stack Overflow for Japanese (SOja), we can use full-width alphanumeric characters for tags. In Japanese, these characters are used almost the same as their half-width version.

Example:

half-width: a
full-width: ａ

half-width: 1
full-width: １

At SOja, we prefer to use half-width characters for tags, and the system also tends to support to use half-width tags. In fact, at SOja, a tag written by full-width alphanumeric characters are usually identified with its half-width version. For example, if we move a mouse cursor on a tag [ｃ], it shows a tag description of a tag [c].

They, however, are not tag-synonyms. We can still tag a question by a full-width one. Searching by a tag is also treating them as different tags.

Example:

• [c] https://ja.stackoverflow.com/questions/tagged/c
• [ｃ] https://ja.stackoverflow.com/questions/tagged/%ef%bd%83

Also, it seems that low-rep users can use a full-width tag if its half-width version exists. For this reason, it often happens that full-width tags are used wrongly.

Currently, SOja users correct these tags by hand, which is time-consuming. Related issues were post to SOja Meta in the past (1, 2), but a SOja manager at that time said "This is difficult to fix. The root cause of this problem is the difference of character handling between SQL and C#".

But it is troublesome to check all tags and detect full-width characters manually. Is it possible to avoid creation of tags with full-width alphanumeric characters? For example, can we restrict these characters (say, U+FF01...U+FF5E) for tags? Or, can we treat them as half-width characters like tag synonyms?