Mapping of fullwidth latin characters to plain ASCII

It would be interesting to automatically map the fullwidth and halfwidth characters to their ASCII counterparts.

"ａｂｃ", "ＡＢＣ" and "１２３" to "abc", "ABC" and "123"…

There are also "（）＜＞［］" to match to "()<>[]"…
There are a lot of characters which could be automatically mapped this way without changing their shape nor signification IMHO.

So many japanese tracks and artists wouldn't be duplicates anymore (without any administration task asked):

少年Ｓ = 少年S
"Ｓ" = "S"

ラッツ＆スター = ラッツ&スター
"＆" = "&"

愛　ありがとう = 愛 ありがとう
"　" = " "

プラットホーム～Merry Go Round～ = プラットホーム~Merry Go Round~
"～" = "~"

I think this would be harmless. Automatic and error free (or so it seems).


PS. Is the search function of this forum really working?

I think this would be a great implementation and something to work on. I too have noticed that this occurs semi-often. It would clear a lot of duplicates up in different languages other than English (genres that I frequent often span over languages).

I am not sure how much this would take. The search function isn't entirely useless, and I see how this implementation could actually help the search function. Searching something and then seeing something nearly exactly like it but with fullwidth characters has happened more than once for me. Usually I go with the option that is not fullwidth, as it normally has more users and is correct.

That is truly a good question.
I wonder if it'd have to manually be gone through and each fullwidth character would be assigned to it's non-fullwidth character. Especially with Asian languages this could get tricky and tedious.

DFA1979 said:
As for the search function: no, it's useless. Just like the search function on every web forum I know :-/
Well, I really like forum search function in general (phpbb, …) but with this one, I can never find anything. The problem is that, when a post is further the second page, it's quite certain to be forgotten for good.

Kerensky97 said:
I'm not a programmer but is this able to be setup to work along the same lines as matching capitalization?
As in how "Track" matches to "track".
Yes, I would think so.

Carla_bruni said:
I wonder if it'd have to manually be gone through and each fullwidth character would be assigned to it's non-fullwidth character. Especially with Asian languages this could get tricky and tedious.
I counted 128 halfwidth asian characters (thanks to babelmap: ｱ｢｣ﾰￗ… U+FF61 ~ U+FFDC, U+FFE8 ~ U+FFEE). I did not think of them at first but yes, why not.

At first, the latin characters would be cool enough and they are only around 2×26 + 10 of them (abc, ABC, 012).