Template talk:ja-romaji

to be discussed

See Wiktionary:Beer_parlour/2013/February#Stripping_extra_info_from_Japanese_romaji. --Anatoli ^{(обсудить}/^вклад) 00:59, 22 February 2013 (UTC)

Category:Japanese romaji without a main entry

I have added this part:

<includeonly>{{#if:{{NAMESPACE}}||<!--
  -->{{#ifexist:{{{1|}}}<!--
  -->|<!--
    do nothing
  -->|<!--
    -->[[Category:Japanese romaji without a main entry]]<!--
  -->}}<!--
-->}}</includeonly><!--

To add entries without an entry it links to Category:Japanese romaji without a main entry as in Template:got-romanization of but it has added extra spaces to romaji entries. Can someone fix it please? --Anatoli ^{(обсудить}/^вклад) 01:06, 4 April 2013 (UTC)

Thanks for fixing, Wyang. It doesn't seem to work well, though. Should it check for all six parameters? It has added many entries to Category:Japanese romaji without a main entry when a kana entry exists (blue linked). --Anatoli ^{(обсудить}/^вклад) 01:13, 4 April 2013 (UTC)

Thanks, ‎Metaknowledge. -age has kana entry and many others. We still have many false positives. --Anatoli ^{(обсудить}/^вклад) 01:31, 4 April 2013 (UTC)

I reverted your last edit. It doesn't matter if it's deprecated; we can remove that when it's actually gone. I'm not sure why some of the false positives still exist, though. —Μετάknowledge^{discuss/deeds} 02:21, 4 April 2013 (UTC)

I was just replacing hira with 2 (unnamed second parameter), which is current. You see, there are 6 unnamed parameters and deprecated hira, hira2, hira3, kata, kata2 and kata3. Checking for all 12 is harder than just 2. Perhaps we should just worry about six current params? --Anatoli ^{(обсудить}/^вклад) 02:26, 4 April 2013 (UTC)

I think we should remove all instances of deprecated params (through AWB, optimally; if you can't do it, make a request at the GP or ask a user like Mglovesfun or -sche) before using this category, because too many #ifexist calls is expensive and will affect performance. —Μετάknowledge^{discuss/deeds} 02:29, 4 April 2013 (UTC)

Mglovesfun opposes this format, AFAIK. -sche was busy with Russian, I asked him before. Besides, when I asked about some help on this template, I only got criticism. I have asked CodeCat, because I borrowed this logic from his Template:got-romanization of, which works (it only has one parameter). --Anatoli ^{(обсудить}/^вклад) 02:33, 4 April 2013 (UTC)

False positives still exist and what's interesting, many of them have hira, a deprecated parameter. Can we remove the check for hira? That way, we can also clean up entries using the deprecated parameter. --Anatoli ^{(обсудить}/^вклад) 02:54, 4 April 2013 (UTC)

For example? Wyang (talk) 02:56, 4 April 2013 (UTC)

Yeah! Thanks for fixing! Wyang, could you check for 6 current unnamed parameters and stop checking for deprecated parameters - hira, hira2, hira3, kata, kata2 and kata3? E.g. abekkuhōmuran uses "kata". Let it be a false positive, so that we clean and remove hira and kata? --Anatoli ^{(обсудить}/^вклад) 03:01, 4 April 2013 (UTC)

Should be done now. Both are categorised into the same category, but labelled differently. When this template is used with both deprecated parameter(s) and nonexistent links, it is categorised into the former. Wyang (talk) 03:12, 4 April 2013 (UTC)

Thanks a bunch! 非常感谢！ What do you mean by labelling? Sorting? --Anatoli ^{(обсудить}/^вклад) 03:24, 4 April 2013 (UTC)

Yes, sorting, if that's the correct terminology :) Wyang (talk) 03:29, 4 April 2013 (UTC)

Wyang, could you please check kaeru, カエル doesn't exist but the entry was not picked up. --Anatoli ^{(обсудить}/^вклад) 04:24, 4 April 2013 (UTC)

Should be fine now. Wyang (talk) 06:47, 4 April 2013 (UTC)

干得好！ --Anatoli ^{(обсудить}/^вклад) 09:30, 4 April 2013 (UTC)

We have cleaned the deprecated parameters but we have some false positives, e.g. zenko in Category:Japanese romaji without a main entry. Could you have a look, please? --Anatoli ^{(обсудить}/^вклад) 11:39, 7 April 2013 (UTC)

It's a server refreshing issue. If you go to edit (the category should be down at the bottom), submit without any change, and then go to edit again, the category probably will disappear. Wyang (talk) 11:46, 7 April 2013 (UTC)

I edited and saved the page without change but didn't see any change. Need your help with this. --Anatoli ^{(обсудить}/^вклад) 12:08, 7 April 2013 (UTC)

I'm currently seeing 50 pages in the category, all of which seem to be correctly categorised: Aoi, aozamezoku, bishōnen, biwa, bui, chanoyu, chū, debu, doji, emu, gen, gokkun, gwa, hadōken, higurashi, hishi, hokku, jin, ken, kizuna, kokku, koma, kora, kotsu, kwa, kyōku, mana, niwatori, Pekin, ratengo, rōmaji, shogi, shuriken, shō, tahibari, taiin, tokudane, Tokushima, tsumajiro, ugui, va, ve, vi, vo, we, wi, wo, yare, zazen, zāzā. Wyang (talk) 12:17, 7 April 2013 (UTC)

I can see 50 entries now too! Not sure what happened before but thank you very much. Do you mind helping me with Template:pinyin reading of, please? It's a similar issue. --Anatoli ^{(обсудить}/^вклад) 12:31, 7 April 2013 (UTC)

RFD

The following information passed a request for deletion.

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

---

Template:ja-romaji

This template is in direct conflict with our policy page WT:ELE. All entries must have an inflection line and a definition introduced by the hash character #. This template attempts to subvert policy without the community consensus needed to achieve such a far-reaching change. Therefore, it should be sent to space and never return. -- Liliana • 21:57, 25 March 2013 (UTC)

The template produces # on a new line with links to hiragana and katakana entries. The way to handle Japanese romaji entries is a result of a long discussion and a consensus with advanced editors who usually work with Japanese who created those entries and edited them. Like Mandarin and Gothic romanization entries, Template:ja-romaji doesn't use inflection lines. The template is used, it's part of the policy for Japanese romaji entries and of course, it should be kept. --Anatoli ^{(обсудить}/^вклад) 22:25, 25 March 2013 (UTC)

• JA romanizations can correspond to one, two, or more JA kana spellings. Requiring multiple invocations on separate # lines seems a bit like tedium for tedium's sake. If a technical policy discussion is warranted, let's discuss. Nominating this for deletion without first discussing the technical issues seems a bit rash. -- Eiríkr Útlendi │ Tala við mig 22:29, 25 March 2013 (UTC)

WT:ELE shows how entries are normally formatted, but I can’t find the part where it says they “must” have a hash character. Can someone help me out with a quote or section link?

I can see advantages in duplicating and improving the same functionality with a template. Anyway, this bears wider discussion to avoid people going off. —Michael Z. 2013-03-25 22:40 z

I can't find this but originally the template had * on a new line, then we changed to #, so it's not picked up as a badly formatted by the bots. A Japanese romaji entry, e.g. "tsuku" is formatted exactly as a model Mandarin pinyin entry "yánlì"

As Eirikr pointed out, romaji serves as a disambiguation to kana spellings, kana spellings in turn link to multiple kanji. Even compared to Mandarin entries, there are more steps (Mandarin pinyin to hanzi - one to many, Japanese romaji - kana - kanji - one to many - one to many). --Anatoli ^{(обсудить}/^вклад) 22:52, 25 March 2013 (UTC)

With the previous system, the ideal was to have the romaji page match a kana page exactly (except for proper nouns which we generally write with the first letter in upper case--so けんじ gets both kenji and Kenji. The way you do that is copy, paste, and tweak the parameters. Copy-and-paste was the ideal. Something about that seems wrong to me. If some editor falls short of the ideal, and adds another Kenji to the page of Kenjis without copy-pasting it to けんじ? There are quite a lot of ways to write Kenji, so I'm sure I missed a few. If the romaji page is added, a dutiful copy-paster has to mirror the change. Personally, after several thousand copy-pastes (usually creations) I began to feel that I was wasting my time. If someone can make a bot to make all the romaji pages, ok, as long as another bot can keep them all in sych with their mirror kana pages. I've never heard of a bot that can merge glosses.

I'm not an expert on all of the policies here, but there's an even more basic principle that a page should have information that belongs on that page and only on that page, that it should have some purpose for existing and not merely be a mirror. Should every romaji page have a disclaimer under "Edit" that says "you may edit, but don't add anything valuable here, it's just a copy of another page" ?

Regardless of the outcome of this kerfuffle, I'm not making any more copy-paste mirror romaji entries. Good luck finding people who will. --Haplology (talk) 02:18, 26 March 2013 (UTC)

You did a great job when you thought it was the right thing to do but yes, romaji entries get out of synch quickly and as you said, it is a waste of time to duplicate the same information when it could be kept at one place. Simplifying the romaji entries do not destroy anything because hiragana entries have that and more. Kanji and kana entries always provide romanisation, including all user examples and inflection, which is a bonus few other languages in non-Roman script here enjoy. It is a very hard job to keep entries in sync, let alone continue doing this for ever. With simplification, more time will be available to work on real Japanese script entries and make sure that all romaji entries are created (this can be done with a bot and won't take long for a human to create, even with less knowledge of Japanese) and no native script variation is missed. I see it as a huge improvement, not a loss. Nothing is destroyed. --Anatoli ^{(обсудить}/^вклад) 02:31, 26 March 2013 (UTC)

@Mzajac: I think you're looking for Wiktionary:ELE#Definitions -- Liliana • 15:37, 26 March 2013 (UTC)

Yes, that’s the bit I was referring to. I don’t infer “all entries must” in that. —Michael Z. 2013-03-26 17:55 z

The result of Template:ja-romaji is also very similar to a huge number of term in different languages - Category:Alternative_forms_by_language with # Alternative spelling of or # Alternative form of, which are not definitions but soft redirects. Template:ja-romaji uses # See: and doesn't allow to add definitions, only new links (up to six, which we have decided to be sufficient for hiragana, katakana variants and mixed scripts). --Anatoli ^{(обсудить}/^вклад) 22:29, 26 March 2013 (UTC)

I dislike the format, it should either be a headword-line template, or a definition template, but not both. Make it analogous to {{hanja form of}}, {{pinyin reading of}} and so on. Mglovesfun (talk) 12:54, 27 March 2013 (UTC)

• To what end, though? The current format supplies both headword and def lines in one convenient package. Splitting it into separate headword and def-line templates doesn't seem at all useful. -- Eiríkr Útlendi │ Tala við mig 17:53, 27 March 2013 (UTC)

Mglovesfun, the template is part of the vote, it was a major point in the BP discussion, was added to WT:AJA#Romaji_entries. Please also see my response on Wiktionary_talk:Votes/pl-2013-03/Japanese_Romaji_romanization_-_format_and_content#Consistency where Eirikr and I tried to clarify the differences. Does your opposition to the template mean that you also oppose the vote Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content, which also includes the format? --Anatoli ^{(обсудить}/^вклад) 22:30, 27 March 2013 (UTC)

I'm suggesting reforming the template, not deleting it. The lack of a hash is very confusing both for human and bot editors; KassadBot (talk • contribs) will be tagging these with {{defn|Japanese}} because of the lack of a definition preceded by a hash. As I human editor I find it confusing too, it makes me want to add a definition or add {{defn|Japanese}}. So in short it makes sense for ease of use for human and bot editors (which is all editors). Mglovesfun (talk) 16:09, 28 March 2013 (UTC)

• @Mglovesfun:

1. KassadBot either isn't running, or isn't flagging these entries. I just checked Category:Japanese_definitions_needed, and none of the romaji entries using {{ja-romaji}} are listed. The only romaji entry currently listed is the decidedly malformed entry at [[rin#Japanese]]. If KassadBot is running, then presumably either 1) the bot doesn't see these as confusing, or 2) Liliana or someone else has reprogrammed the bot to recognize these entries as okay.
2. As a human editor, if you saw an entry in the regular rendered view that appeared correctly, why would you feel any need to change the wikitext? I run across various bits of wikicode that do things I don't recognize all the time; maybe I'll look at the template source or relevant Help pages if I'm curious, but so long as it renders correctly, I generally leave it be.

• As a bit of background, I believe that part of why Anatoli decided on the original structure of the {{ja-romaji}} template is that it more closely mapped to the POS headword templates previously in use, such as {{ja-noun|hira=XXX|kata=YYY}}. This can be turned into a working call to {{ja-romaji}} simply by replacing noun with romaji, and then deleting all JA content below. In terms of ease of use, this is extremely easy indeed; splitting this into two separate templates, one for headword and one for def lines, and worse yet requiring one template call per def line, would be significantly less easy.

Later on, the decision was made to forgo labeling links as "hiragana" or "katakana" and just say "See:", obviating any need for named params, so I reworked the template to use just positional params. However, I kept the named params for now as deprecated, so these still work. Eventually we will remove the named params from all template calls, and then remove them from the template.

I'm operating from the working assumption that templates exist to make wikicode easier to produce and work with, faster to enter, and shorter. Requiring that a working template be split up into multiple different templates and multiple separate template calls, all for the sake of longer wikicode but identically rendered HTML, runs counter to the whole point of having templates in the first place. But then perhaps my working assumption for why to use templates doesn't match that of other editors? -- Eiríkr Útlendi │ Tala við mig 17:47, 28 March 2013 (UTC)

KassadBot is running but it won't pick up romaji entries as badly formatted from the moment "*" was changed to "#" (on a new line). I find the template a big improvement and it's very easy to use. --Anatoli ^{(обсудить}/^вклад) 05:25, 29 March 2013 (UTC)

It must have a special exception for Japanese. Otherwise, I don't know. -- Liliana • 20:49, 4 April 2013 (UTC)

To answer to your comment I need to know your motivation. Are you against limitation of the Japanese romaji entries or against the way definitions are done or both? The rationale and the technical solution was discussed for a month and agreed on by JA editors on BP. Well, templates are for doing what people would do manually otherwise. Nothing is broken. The definition line is there as per Wiktionary:ELE#Definitions, only it's generated by the template to make sure the format of romaji entries is followed. Consider this an exception if you wish. Japanese romaji have been an exception for a while. You are welcome to vote against both romaji simplification and the template in votes Dan Polansky has created. The template is now documented - Template:ja-romaji/doc. Also, see Wiktionary:Beer_parlour/2013/April#Gothic_romanisation_template for a similar approach for Gothic. --Anatoli ^{(обсудить}/^вклад) 21:52, 4 April 2013 (UTC)

• @Liliana, I suspect that KassadBot is looking at something closer to the final rendered version, perhaps after template expansion. Since the template does provide for individual def lines, each starting with a #, this would presumably pass KassadBot's inspection. -- Eiríkr Útlendi │ Tala við mig 22:36, 4 April 2013 (UTC)

It does pass KassadBot's inspection. --Anatoli ^{(обсудить}/^вклад) 22:40, 4 April 2013 (UTC)

Having overlooked this deletion proposal, I have created a vote related to it: Wiktionary:Votes/pl-2013-03/Romanization and definition line. The vote does not propose to delete ja-romaji template, but it proposes that we require each entry to have a definition line in the wikitext. --Dan Polansky (talk) 08:30, 30 March 2013 (UTC)

I have detagged the template. It's used in many entries now. --Anatoli ^{(обсудить}/^вклад) 23:35, 1 April 2013 (UTC)

RFD-passed. —Μετάknowledge^{discuss/deeds} 17:41, 19 July 2013 (UTC)