[eesmith]

My I suggest the options, like "Hide non-matches", be at the top? Then they are easier to spot, and are in a persistent location relative to the input area, even if characters are hidden.

I like on my Mac how its choose shows similar characters, so "u" suggests "U", "ü", "ú", and so on.

[reknih]

I put the buttons on the bottom so the user would not be overwhelmed by options at the start. Interesting to hear you'd prefer them on top of the table!

Suggesting similar characters would definitely a nice feature, I'll add it for basic cases.

[eesmith]

I found a neat oddity. I was browsing through the code pages and saw a "P" with a horizontal bar through it, in the OEM Nordic cp=865 page at 0x9E. I wondered what it could be.

It is "₧", which is U+20A7. https://en.wikipedia.org/wiki/Spanish_peseta comments:

> In the version 1.0 of Unicode the character ₧ U+20A7 PESETA SIGN had two reference glyphs: a "Pts" ligature glyph as in IBM code page 437 and an erroneous P with stroke. In Unicode 2.0 the reference glyph P with stroke was erroneously displayed as the only symbol for peseta[5] and was later corrected to the Pts ligature and a separate character code was added for the peso sign.

pointing to why the P-with-bar unification of the Spanish and Filipino peseta doesn't work, at https://www.unicode.org/L2/L2000/00013-n2156.pdf .

The visual display seems to come from the use of Grotesk, see https://fonts.google.com/?preview.text=kl%E2%82%A7&preview.t... for a couple examples with that presentation.