I personally prefer using the original meanings of character components because they (may) provide clues to pronunciation and meaning. This also means I have to look up the components first.
In that regard I noticed that you’re using images instead of characters, e.g. instead of 䜌 or 龻.
In this particular case it seems most characters use the reading (luán) of the component, so whatever it may resemble was never that useful anyway. This was quite eye-opening to me actually. But I had to work backwards from already knowing 灣 because of the image thing. So bit of a chicken-and-egg problem
Somewhat of a power-user feature I guess. I hope you’re not spending an awful lot of time photoshopping characters, because most components do have real unicode entries
Some of the images are because the font we use (
Noto TC) does not include some of the unicode characters. For your example:
There is no glyph for that unicode entry. We opt to just using images instead of defaulting to another font, since some of the even rarer glyphs are not only not within our chosen font, but also not within the system font either !
One thing I’ve thought about but will likely not get to anytime soon is to edit the font file itself so we get our own proper glyph for some of these, allowing us to avoid the images most of the time. Hell, for some of the ones that have no real unicode entry, we could even use the Unicode Private Use Areas with our own glyphs as well, if we were feeling adventurous.
Oh, I didn’t realize the complexity of the problem!. Classic case of xkcd: Tasks