掲示板 Forums - [Discussion] radical schedules + counters

I don't understand what this is trying to say:

"We do not have enough data do have a counter selector ( 4 + magazines = よんさつ), but this is possibly something we could crowd source."\

Could you elaborate ?

As for counters:

I think a vector that does something very similar to Counter punch i.e. show you a picture and then expect you to answer with the correct term e.g. １０本, would be nice.

Picture -> Hiragana

Picture -> Kanji

With options of multiple choice and fill in the blank

Radicals in kanji are far from ideal. First off, the UI is a mess - radicals don't have readings to be concerned of, stroke count is not really needed, and then a radical marker of a radical? There's be "related words", but of course, no related words.

The majority of the study vectors and quiz modes for kanji schedules would not work for them, and adding special exceptions for all of this inside of the kanji schedule (or kanji quiz settings) would add a ton of complexity.

Separating them, though, gives a much simpler overview, especially for those that aren't power users. In addition, we can customize the info boxes for radicals to better enhance them. For example, instead of related words, we can give a few examples of kanji with those radicals, maybe even giving preference (if possible) to ones that have meanings that are somewhat linked to the radical itself.

Additionally, thanks to our amazing (and growing) mnemonic database, it would be relatively trivial to say something like:

Often appears in mnemonics as: X, Y, Z, showing other meanings that are frequently co-opted by other users in mnemonic creation.

A significant number of users have asked for this in the past, with one of the key reasons being wanting a way to make the user mnemonics more accessible.

As to the counter thing; the "item" data from Counter punch is extremely limited - maybe 4-5 nouns per counter, and that is only for the ones that Counter Punch handles. I'd much rather have a larger set of 50+ counters, each with 10-20 different words that work for them. That would provide a much deeper grounding in each counter, and make the quizzes much less repetitive. Instead of relying on Saki's limited set of drawn pictures, we could rely on the much larger set of vocabulary imagery.

It could also allow us some more interesting quizzes, such as "groupings" (similar to the "match the reading" quiz mode) where you have 3-6 counters, maybe 10 words, and you have to drag them to the right counter.

I wonder if there is a publicly accessible (=usable) list of counter + words used that we could use as an initial pool...

Thanks for the explanation. That makes sense.

Given that, in addition to the "number + noun -> hiragana" vector, (the 4 + magazines -> よんさつ questions), which covers how it's read, it would be nice to have an additional vector:

number + noun -> number + kanji counter, e.g. 4 + magazines -> 4冊

Which would cover how to produce the common written form