Re: [edict-jmdict] More radical thoughts

[Jim Rose <jim@*************>]

OK Cool... whatever you decide is fine with me - and in fact a bit relieved to leave it up to you frankly.   Just wanted to make absolutely sure you don't want any of those in there.  Now I can go back to finishing the code I started instead of fretting over missing radicals should be there or not.  Next public project I start should be a radkfile redo maybe... perhaps using some a statistical approach to "official" decompositions, if enough people were willing to decompose a few thousand kanji in their spare time to make statistics matter.  Doubt that many people will participate... but who knows.

On Nov 6, 2006, at 9:45 PM, Jim Breen wrote:

[Jim Rose ([edict-jmdict] More radical thoughts) writes:]
>>
>> KANJIDIC also list these as radicals and or variants of radicals not
>> currently displayed in radkfile derivatives (most notable are 巛
>> and 艸

Correct. As I said earlier, the decomposition in radkfile/kradfile and
the 214 classical radicals and their derivatives are not the same things.
radkfile makes major use of things like マ and ム which are not in the
214.

[...]

>> Just wondering where we could substitute, where we could add.

With 251 elements in use at present, I am not at all keen to add any more.
If more are added, it will be because they a identifiable visual elements
that are istinct and common enough to justify their inclusion; not because
they happen to be on lists of radicals. I don't know of too many kanji
with obvious 艸 patterns. How many have you found containing JIS212 69-93
U9578?

Substitution is easy, but I think all the obvious ones have been done.

Cheers

Jim