[unihan] Unihan variants information

[jenkins via Unicode]

Very impressive! Thank you for this.

> On Jan 17, 2020, at 6:03 AM, Michel Mariani via Unihan <unihan at unicode.org> wrote:
> 
> FYI, the "Unihan Variants" utility has been recently added to the open-source application Unicopedia Plus <https://github.com/tonton-pixel/unicopedia-plus>.
> It provides both the linear and structured informations planned about one year ago.
> I think that the graph view available in SVG format can be especially useful to spot possible inconsistencies between variant properties...
> HTH,
> 
> 	--Michel MARIANI
> 
> <unihan-variants-turtle-screenshot.png>
> 
>> I've developed an open-source, multi-platform desktop application called Unicode Plus <https://github.com/tonton-pixel/unicode-plus>, which is a set of utilities related to Unicode, Unihan and emoji.
>> 
>> The basic Unihan-related utilities are almost completed, and now I would like to add more useful information about the Unihan variants:
>> 
>> 1. First option: "Linear Information"
>> 
>> - A linear list of all the variants *related* to one given Unihan character would be displayed, similar to what can be found in Apple's Character Viewer (or Palette), or in the "Unihan Variant Dictionary" application.
>> 
>> - Two sources of data could be merged:
>> 
>> 	1. The information provided by the "Variants table for Unicode" data file UniVariants.txt <http://kanji.zinbun.kyoto-u.ac.jp/~yasuoka/ftp/CJKtable/UniVariants.Z> by Prof. Kōichi Yasuoka.
>> 	
>> 	2. The information extracted from the relevant Unihan DB tag properties: kSemanticVariant, kSimplifiedVariant, kSpecializedSemanticVariant, kTraditionalVariant, kZVariant.
>> 
>> - Discarding self-variants, assuming that Z-variants are somehow symmetrical, and possibly merge the different types of variants tags would result into independant sets of *related* Unihan characters. Accessing the info would then simply imply testing which set a given character belongs to, and omit the character itself for display.
>> 
>> - This kind of information is most certainly user-friendly, however it lacks structural information about the relationships between the different variants.
>> 
>> 2. Second option: "Structured Information"
>> 
>> - This is probably more ambitious and challenging: ideally, the information could be displayed graphically as a diagram of characters joined by arrowed links, indicating the type of variant. It would support one-to-one, one-to-many and many-to-one relationships...
>> 
>> 
>> Any ideas, comments, suggestions are most welcome...
>> 
>> -- Michel MARIANI
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://unicode.org/pipermail/unicode/attachments/20200117/8c65fe24/attachment.html>