Module talk:ja-see/old

Improvements for categorization

@Dine2016: First of all, great job on creating this module. The {{ja-see}} template has a clean interface and reduces content duplication across entries.

Anyway, I would like to suggest the following improvements for categorization:

1. Automatically add hiragana entries into Category:Japanese hiragana even if the kanji entries are not yet created. For example, しゃっかん currently does not belong to any category because 借款#Japanese is not yet created.
2. Suppress categorization of hidden categories such as Category:Reference templates lacking the author or editor parameters, Category:Foreign word of the day archive, Category:Requests for attention concerning Japanese, etc which are not relevant to the hiragana or alternative kanji entries. KevinUp (talk) 16:40, 16 September 2019 (UTC)

@KevinUp: Hi. Thanks for bringing up the issue. Regarding the second point, do you happen to be aware of any other categories that need to be excluded from copying? I confess I'm not familiar with Wiktionary's maintenance system, so any help would be appreciated.

The reason I kept the interface of {{ja-see}} clean was because I was not good at programming (scripting), so I wanted to make the current implementation easily replacable. --Dine2016 (talk) 04:18, 17 September 2019 (UTC)

@Dine2016: Thanks for adding categorization for the hiragana entries. I think most of the categories that have been copied over are fine, except for hidden categories such as Category:Sort key tracking/redundant and other categories I mentioned above. Is there a way to exclude these categories from getting copied over? (The hidden categories can be seen by enabling "Show hidden categories" under the "Appearance" tab of "User preferences"). If it's not possible to automatically exclude these hidden categories, can you add a line in the code to manually exclude these categories from getting copied over?

(By the way, this is a very good module and may be useful to reduce content duplication in other languages such as Javanese etc. which have same terms in different scripts). KevinUp (talk) 20:00, 17 September 2019 (UTC)

@Dine2016 I noticed that あぢ, やう, 渦巻線 are not being categorized. Any idea what's going on? Also, the definition for なにいろ is not getting copied over. Please check, thanks. KevinUp (talk) 09:55, 28 September 2019 (UTC)

@KevinUp:

1. {{ja-see}} does not support double redirects, so あぢ should point to 味 and 鯵 directly. Technically, it is possible to let the editor just specify {{ja-see|あじ}} and make the template find the lemmas, but I didn't implement it that way because it would consume more memory.
2. It's best to create the lemma entry first and then alternative spellings. So 渦巻線 should be created after うずまき線. The reason I didn't give alternative kanji spellings a category is because they would require {{ja-kanjitab}} anyway.
3. When there are several words on the lemma entry, they should be separated by etymology headers on the L3 level. 何色 is lacking two etymology headers.

I recognize that the presumptions above is not clearly stated in the documentation of {{ja-see}} due to my inability to write good English. I will try to make this clearer later. --Dine2016 (talk) 14:22, 28 September 2019 (UTC)

@Dine2016: Thanks for replying and fixing the entries at あぢ, 何色, 渦巻線. Now I know what's going on. Actually, your writing ability in English is good. I've updated the documentation of {{ja-see}} based on your replies above. KevinUp (talk) 09:16, 29 September 2019 (UTC)

@KevinUp: Thanks. Actually, much of the trouble arose from the lack of a one-to-one correspondence between Wiktionary entries and actual "words". If Wiktionary used homograph numbers in page titles (so that one entry corresponds to one word) then the "etymology headers on the L3 level" requirement would be unnecessary. If Wiktionary went a step further and used full titles like あじ, あぢ, 味 (so that one word correspond to one entry) with proper searching functions then even {{ja-see}} would be unnecessary. Unfortunately, {{ja-see}} at the current stage had to use etymology headers to distinguish between different "words" and requires the lemma entry to list alternative spellings in specific places so that it knows which "words" to copy. --Dine2016 (talk) 12:53, 29 September 2019 (UTC)

I think this is not easy to implement because most users would enter the word directly into the URL to find an entry (I've even come up with a command-line version for a less cluttered appearance). A wikidata-like approach using numbers to distinguish between different "words" might be possible, but the two projects are disconnected. For now, I'm pleased with what this template is able to do. KevinUp (talk) 18:30, 29 September 2019 (UTC)