If you know the word for window (窗户), perhaps (也许) and night (夜). PS idek if these are the right hanzi for the words
Then the Mneomic would change to OPTION A:
([r-] Robin Hood) is camping in the ([4] bathroom) of the ([-_] house), shooting arrows at the (日 sun) through the open 窗户, which of course never reach. He wants to steal the (日 day) from the (日 sun) and let 夜 owls have their reign, 也许.
I know the grammar wouldn’t be perfect 1:1, and some words may not match 1:1 via translation but think this would be beneficial to people getting some active practice reading the words in a more “normal” context.
Note, I did not put the english translation next to the word like OPTION B:
([r-] Robin Hood) is camping in the ([4] bathroom) of the ([-_] house), shooting arrows at the (日 sun) through the open (窗户 window), which of course never reach. He wants to steal the (日 day) from the (日 sun) and let (夜 night) owls have their reign, (也许 perhaps).
This is to make it more natural to rely on what we learned, maybe if possible have two different toggles/options for Option A and Option B. For Option A you can enable hovering over it to show the translation vs showing it directly.
How to do this? Not sure, depends on how the mnemonics are currently constructed. If it’s currently just text, an inefficient way off the top of my head is each mnemonic can be turned into a map/dictionary and tested against someones already learned words. A fun engineering problem
I remember seeing a browser extension that did this for WaniKani (a Japanese Kanji learning app that has some similarities to HanziHero). The benefit of that approach is that it can be used when browsing any web page, if desired.
I think when we get around to exposing a read-only API this should be easy to implement as web extension if anyone is interested in doing so. Probably just a map of English meanings to HanziHero characters or words that have been learned by the user already. Since there is a large amount of overlap, it would need to select a random out of the suitable matches. It would probably also want to ignore alternate meanings and only match if the exact primary meaning matches (maybe stripping out the "to " if it exists in the primary meaning to facilitate this).
I would be curious to hear others thoughts on the utility of it. It’s a nice trick to get some more meaning reinforcement, but I’m unsure of how useful it is in practice.
I found a browser extension that did something similar this. The idea is that it slowly ramps up the number of words its replacing over time so at first it’s like 1 word a paragraph and soon it’s like 80% your target language. It just went on frequency list rather than your own list of known words, but the idea was the same. It sounded cool and so I tried it. Isn’t that great you get to read your normal stuff but also practise your target language with no effort, yay …
… and I lasted about 3 days I think before deleting it. There just isn’t a real way to make it work except for maybe the simplest of nouns. Translating individual words into your target language just has waaay to many flaws and pitfalls (i.e. you can’t replace and with 和 even if the English translation is an exact match etc). Especially once you get above 1 word per sentence. The context + word placement is just all wrong! It makes this weird broken hybrid pidgin where it sort of makes sense but also is really wrong at the same time and I figured it was almost doing more harm than good (it usually made the English more confusing rather than making the Chines characters easier to remember).
However if you have a person manually select what words to translate in which sentences so it made sense I think it’d probably actually be great (like Nedap did above). So if you’re volunteering to re-write all 387844 mnemonics sounds great
