Radical definitions incomplete?

[tilusnet]

Hi,

I noticed some radicals don't come up with definitions in the tool, eg. radicals 66, 51, 52, etc.
Are these meant to be defined inside the definitions.json files, or organised differently?

[mreichhoff]

ideally they'd exist in that file, yes, or be in the partitioned definitions data (I should probably do away with the partitioning for definitions and just load the whole dictionary on startup...with brotli it's not that large a file to load over the network, but that's another task for another day).

I'm not super familiar with the radical numbering scheme, but wikipedia seems to say this is kangxi radical 66, and it has definitions? https://hanzigraph.com/simplified/%E6%94%B4

Wanna send the specific radicals you weren't finding in text form? I can then check whether CEDICT has definitions, and if so, figure out what bug I have.

[tilusnet]

Yes, I find the coverage of radicals particularly useful in the Tree view where you decompose characters.

For instance, take 邀, once I "drill down" to ⺙, (see image) which is radical 66, I get no definition for it.

In contrast, if one does a similar "drill down" on MDBG for the same word, you get the definition of radical 66 as expected:

[tilusnet]

To confirm the radical system, yes, I meant Kangxi.
I'm aware there are variations to these - my understanding however is that Kangxi is the most widespread and considered a "de facto" standard.

[tilusnet]

Covering all radical forms in this table would be great:

https://en.wikipedia.org/wiki/Kangxi_radicals#Table

[mreichhoff]

got it, I suspect the issue is a mismatch of the characters between CEDICT and the components dataset I used: https://github.com/cjkvi/cjkvi-ids

I will see about correcting this

[mreichhoff]

this should be the list of components without a definition in either simplified or traditional data:

definitionless-components.txt

I'm not sure I see a better way than manual or AI classification to find similar glyphs that do have definitions. This might be a challenge, given things like: ⺙ vs 攵 vs 夊 (multiple of these have definitions). I will also dig into what sources mdbg is using.

[mreichhoff]

also of note: HanziCraft's page for ⺙ also lacks definitions: https://hanzicraft.com/character/%E2%BA%99