[Uim] Korean input

[Park Jae-hyeon]

David Oftedal <david at start.no> writes:

> However, that isn't true for this input method, since it doesn't
> distinguish between initials and finals. For instance, the sequence
> "lalala" can become either "라라라" or "랄알아" depending on how the
> user uses the commit key. Or what if the user wants to type "감ㅤㅅㅏㅎ압니다"
> instead of "감사합니다"? It doesn't have to be space, of course, but it
> has to be a key that the user can press fairly often without undue
> stress being put on the fingers.

Ok, now I realize why HWP assigns "x" to leading "ㅇ".
You are better aware of romaja input of Korean than I.  :)

Your two examples are both related to leading "ㅇ".  So, in HWP,
"라라라" corresponds to "lalala", and "랄알아" to "lalxalxa".  Note
that Koreans usually regard leading ㅇ as a consonant, although it has
no phonetic value.  The other example can also be entered using "x".
In this way HWP implements romaja input without a commit key.
However, as I said before, ambiguity still remains, although it occurs
very rarely.  An example is "가까" and "각가".  Both are mapped to
"gagga".  So they should be distinguished, for example, as "ga'gga"
and "gag'ga".

It is true that we need a commit key, and SPACE certainly is a choice.

> >ㅟ  wi
> >ㅢ  ui, eui
> >
> Wow. It might be difficult to include all of these, since each
> additional one requires around 884 extra lines in the table file. I've
> made a script to fix the r/l issue, though, which will be corrected in
> the next version, and I'll try to add some of the other ones. I'm a bit
> puzzled by ㅢ and ㅟ though - Wouldn't it be more correct to map "ui" to
> ㅟ and map the other to "eui"?

The standard romanization states that "ㅟ" should be mapped to "wi",
and "ㅢ" should be mapped to "ui".  I also doubt whether "ui" is an
appropriate romanization of "ㅢ" or not.
See http://www.hangeul.or.kr/24.htm .

Jae-hyeon