Wiktionary talk:Votes/pl-2013-03/Romanization and definition line

Why?

Is not answered in the proposal. —Michael Z. 2013-03-31 02:06 z

To make the job of spotting missing definitions easier for humans and bots; to achieve uniformity of formatting of entries that direct the reader to other entries for definitions, as is the case with inflected form entries and alternative form entries. --Dan Polansky (talk) 13:37, 31 March 2013 (UTC)

Romanization entries have no definitions... -- Eiríkr Útlendi │ Tala við mig 18:19, 1 April 2013 (UTC)

We already have a vote - Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content, which is also about the format. I don't see why we need two votes. --Anatoli ^{(обсудить}/^вклад) 23:30, 1 April 2013 (UTC)

• It appears that this vote was forgotten. Liliana has brought it back up. I think this was effectively superseded by Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content, but I've voted here anyway just in case. -- Eiríkr Útlendi │ Tala við mig 19:29, 1 July 2013 (UTC)

  This vote was definitely was not superseded by the sabotaged vote Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content. This vote raises an issue that was left out of Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content. --Dan Polansky (talk) 20:23, 1 July 2013 (UTC)

  • I think sabotaged isn't the right descriptor. That vote ran its course, with enough community opposition to even having the vote that the consensus that emerged appears to be, “we should let romaji entry conventions evolve according to discussion, not according to hard-binding policy votes.” That would seem to 1) supersede this vote, and 2) indicate that JA editors should deal with JA romanization entries. -- Eiríkr Útlendi │ Tala við mig 20:43, 1 July 2013 (UTC)
    • There was not enough community opposition to even having the vote. The vote was actually sent to RFDO, which failed to lead to its deletion for the lack of support of that action. Of the people who posted to the vote page itself, four voted for opposing even having the vote, while seven chose other section of the vote. You are in a minority, albeit vocal and aggressive one. --Dan Polansky (talk) 20:51, 1 July 2013 (UTC)
    • Re: 'consensus that emerged appears to be, “we should let romaji entry conventions evolve according to discussion, not according to hard-binding policy votes.”': No such consensus can be detected. No such proposal was sent to vote. --Dan Polansky (talk) 20:54, 1 July 2013 (UTC)

I would actually support that vote because it does not say anything about implementation details, just the entry format (which I can support). I oppose the use of {{ja-romaji}} to implement it. —CodeCat 20:59, 1 July 2013 (UTC)

Key difference between JA and CMN use cases

One key difference to note between the sample Japanese and Mandarin examples is that Japanese romanization entries have no gloss. Any gloss would be given on the targeted kana page, while full definitions would be given on the kanji page (if it exists, or on the kana page if there is no kanji spelling). Consequently, for the romanization entries, there is no need for line-by-line definitions, as there are no definitions given. -- Eiríkr Útlendi │ Tala við mig 19:30, 1 April 2013 (UTC)

But the entry does display one or more “definition lines,” as we call them. Being consistent in wikitext with the other 99.99% of entries, to help editors understand it, is a valid argument for retaining a two-template version.

The addition of a Gothic example would help put this in its fuller context. —Michael Z. 2013-04-01 19:55 z

Gothic: wisan, wulfs, þuk. Mandarin pinyin entries should only have links to Chinese characters, no definitions either and Dan Polansky knows is because he was complaining, see Wiktionary:Votes/2011-07/Pinyin entries. The model entry is "yánlì", not the one given in this vote. Quote: That a pinyin entry have only the modicum of information needed to allow readers to get to a traditional-characters or simplified-characters entry; see [[yánlì]] for an example. The template "Template:ja-romaji" ensures that the strict new rules about the structure of the romaji entries are actually followed. Pinyin entry structure was not strictly followed. Any new pinyin entry I make won't have any definition, just links to hanzi, as per vote.

@Mzajac. BTW, Wiktionary:Gothic transliteration uses the wording "soft redirects", which is what the romanisation entries really are (or "indices"), they are not a replacement for native scripts, alternative spellings or common misspellings. --Anatoli ^{(обсудить}/^вклад) 23:27, 1 April 2013 (UTC)

I get all that, except that index and redirect are mutually exclusive concepts. One is for choosing and the other is for not. If we are to treat the three languages’ romanization entries together, then it can’t be as redirects. —Michael Z. 2013-04-02 15:13 z

We don't have to treat romanisation entries for all languages together, the situation is different and different people work with different languages. What is good for one group is hated by another. A similar approach would work for Gothic and most languages where romanisation entries are helpful. All that's needed is a link to a standard spelling (or spellings). Pinyin is different. There are many homophones with the same toned pinyin. The role of disambiguation is taken by hiragana entries in Japanese. There is no need for a definition line in romaji entries, it's generated by the template. Of course, people who don't work with Japanese may not understand why there is no definition line but like any template, it may need some documentation or a guide. This template is very simple to use, though. --Anatoli ^{(обсудить}/^вклад) 22:11, 2 April 2013 (UTC)

Are you saying that romaji and Gothic romanizations are very similar and would benefit from the same approach? —Michael Z. 2013-04-03 14:02 z

Gothic would work perfectly with a similar template (Template:ja-romaji) and with less parameters. I also answered this in Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content today. --Anatoli ^{(обсудить}/^вклад) 22:50, 3 April 2013 (UTC)

• I would agree that Gothic could benefit from similar treatment -- doing so would result in less wikicode, fewer template calls, and smaller page sizes, all of which I view as positives. -- Eiríkr Útlendi │ Tala við mig 17:52, 5 April 2013 (UTC)

It would work not only for romanization entries but in any type of entries where a strict and limited structure of entry definitions is desirable or required, e.g. alternative spellings of, common misspelling of, (any inflected) form of, etc. The definition line(s) would be generated by a template. --Anatoli ^{(обсудить}/^вклад) 00:29, 8 April 2013 (UTC)

Re: Gothic (proof of concept):

I have created Template:got-romanization (different from Template:got-romanization of!) and a sample entry "afdrausjan" (modified to use). As with Japanese Template:ja-romaji the definition line is generated by the template. It has the same look and feel as a new romaji entry. --Anatoli ^{(обсудить}/^вклад) 03:34, 4 April 2013 (UTC)

More discussion

There's more discussion about the whys and wherefores of this template at [[Wiktionary:Requests_for_deletion/Others#Template:ja-romaji]]. Please read that thread for more background. -- Eiríkr Útlendi │ Tala við mig 23:35, 1 April 2013 (UTC)

This vote is superfluous

The status quo already requires putting # before each definition. WT:ELE even says this. So why do we need a vote to confirm what is already policy and common practice? —CodeCat 21:40, 5 April 2013 (UTC)

• Common practice, yes.
• Explicitly stated requirement in WT:ELE, no.

The closest I see to "requiring" this is in item 4 of [[Wiktionary:ELE#A_very_simple_example]] (bolding mine):

This is a simple entry for the word bed, and shows the most fundamental elements of an entry:

1. a word’s language (as a level 2 heading),
2. its part of speech or “type” (as a level 3 heading),
3. the inflection word itself (using the correct Part of Speech template or the word in bold letters),
4. a definition (preceded by “#”, which causes automatic numbering),
5. links in the definition for key words,
6. “References” (as a level 3 heading), and
7. a verifiable place where you found the word

Nowhere that I can find does WT:ELE say that the # is required in the wikitext. The implied requirement of item 4 in the above quote is that entries should list definitions on individual lines, and that those definitions should be automatically numbered.

{{ja-romaji}} does this. As such, current practice using {{ja-romaji}} is in no way counter to the requirements of WT:ELE. -- Eiríkr Útlendi │ Tala við mig 21:56, 5 April 2013 (UTC)

KassadBot and romaji entries

KassadBot started picking on romaji entries form 7th of April and adding notice on missing definitions, even though it didn't do it before. The definition lines with # are there, they are generated by the template.

If this action is going to continue, I will boycott this vote. Category:Japanese_definitions_needed currently contains a few hundred romaji entries, mass-tagged by KassadBot. --Anatoli ^{(обсудить}/^вклад) 05:37, 9 April 2013 (UTC)

I blocked the bot for one week when I noticed it was making that change, which I assumed was due to an error. Subsequently, Liliana wrote that she didn't understand why the bot abruptly started making the change, which doesn't inspire confidence. Given that the format of romaji entries is subject to two ongoing votes and much BP discussion, and the acceptability of {{ja-romaji}} per current WT:ELE is disputed, I agree with Eirikr's assessment that having a bot mass-make chanes to the entries to support one POV "could be viewable as either a willful nuisance or an attempt at strong-arming"—even though I agree with CodeCat and Liliana that current WT:ELE probably (and longstanding practice certainly) does require an explicit #.

I expect that the blot will not resume making the controversial change before discussion and voting has concluded, or will be blocked again if it does. - -sche (discuss) 09:06, 9 April 2013 (UTC)

I'm surprised that Liliana doesn't know, because I remember somewhere else that this very thing was expected to happen. While it's true that the hash sign is in the template, the bot can't see this because it only see the page's wikitext, like normal editors do. To figure out whether the bot includes the proper wikitext would mean expanding the template, which would slow down the bot quite a bit, and it would make things very complicated for the bot. —CodeCat 12:58, 9 April 2013 (UTC)

• @CodeCat, the bot wouldn't ever have to expand the templates on its own. The API has a parameter for fetching expanded content. It seems to be inexpensive for smaller pages, which most romanized pages are; at any rate, I can see no noticeable change in load times comparing a fetch of unexpanded wikitext versus a fetch of expanded wikitext. (You will have to manually hit the Make Request button at the bottom to see the results.) This is for the small entry at [[sasu]], to show the before and after for {{ja-romaji}}. Even for medium-sized pages, I see no major difference in load times: Wiktionary's Main Page unexpanded, expanded. It's not until we look at really complicated pages that load times begin to be a factor, such as the entry at [[a]] unexpanded, expanded.

Since the API offers this expansion feature, and since KassadBot stopped flagging {{ja-romaji}} entries after Anatoli edited the template on 16 March to use # instead of *, and since KassadBot only started flagging this new format for {{ja-romaji}} entries on 7 April despite no change to template format, simple logic leads us to the conclusion that something in KassadBot must have changed. Either that change was that KassadBot was not running from the time Anatoli changed the format until 7 April; or that change was that KassadBot started fetching unexpanded wikitext instead of expanded; or that change was in some other library or piece of code that was altered, and that is external to KassadBot but called by the bot. -- Eiríkr Útlendi │ Tala við mig 18:51, 9 April 2013 (UTC)

At this point

I would support this proposal. - -sche (discuss) 03:28, 27 May 2013 (UTC)