The components feature is completely wrong

[ortiofio]

It decomposes characters by largest stroke parts that it can find are some other characters, while completely disregarding 六書 or real relation between symbols. This is not how Hanzi are formed.

Compare the website's breakdown of 說 and Wiktionary. 言 is unrelated to 亠. 兌 is unrelated to 兄. No mention of what is semantic and what is phonetic.

I understand this could work for someone, if they don't care about how Hanzi actually work, but then there should at least be a warning or clarification of this.

[mreichhoff]

Thanks for the feedback! You're certainly right that the components feature isn't intended to give etymological information. It is a good feature idea to do so. I don't agree that a warning is needed, though.

言 is unrelated to 亠

Pleco and HanziCraft (screenshots below) both indicate otherwise. I think you may want to give this feedback to the data repositories that I suspect back several such tools, including HanziGraph: https://github.com/cjkvi/cjkvi-ids and https://github.com/amake/cjk-decomp.