Special characters like × (ばつ) or ○ (まる) are not parsed

[Gnurou]

These characters (as well as others, like the romanji alphabet) should also be used as search parameters as they are recognized as kanjis in the database. However, the word parsing regexp discards them and it looks like they are not included in the character recognized as kanjis by the software.

[Tracy Poff (tracy-poff)]

A quick search on wwwjdic shows these (× and ○) very little used in actual words. I have seen them used to replace a syllable in e.g. the name of a real-life anime, when referenced in manga or games. Perhaps they should be treated as wildcards, in order to facilitate finding the word that was intended? Of course that being successful would depend on the word with the replaced syllable being a real word, and not something like 'Gundam', but I guess you see my point.

[Gnurou (gnurou)]

I think interpreting them "as is" is the best behavior for these characters. It's actually just a regexp limitation that will probably fix itself when I write the new command system. Anyway as you pointed out this is not really a critical issue.