Wiktionary talk:Votes/pl-2013-03/Japanese Romaji romanization - format and content

Unnecessary vote

I would just like to register my annoyance that after consensus was achieved among all our active Japanese editors, Dan decided to create a vote which does not even match the result of the consensus exactly, seemingly so that we can inject more bureaucracy into the process. —Μετάknowledge^{discuss/deeds} 15:48, 23 March 2013 (UTC)

• Ditto that. Dan, I respect what I see as your view that due process must be followed. That said, consensus has been reached for all the JA editors currently active. Moreover, this change affects only the duplication of information, not the presence of that information itself -- rom entries for Japanese before this change were only ever intended to be duplications of content already on the corresponding kana entries, and were provided entirely for the sake of the convenience of users who cannot yet read kana.

Maintaining duplicate data sets has proved to be onerous enough that Haplology and Anatoli wanted to move away from this practice; after working out my own concerns about usability and access, I acquiesced and have been actively supporting the change by reworking rom entries. I hope to soon be able to create a bot to handle this. I had originally started reading up on bot creation in order to create and maintain the rom dupes of kana entries; really, going either way (i.e. creating rom dupes of kana, or creating rom stubs as soft redirects to kana) is theoretically very simple, and is much more a task for automation than for human editors.

Anyway, to restate, this change in the handling of romaji entries does not involve the removal of information from EN WT, and instead does involve the collapse of duplicates into one single location. This could be viewed more as data normalization. And since that data still exists, we can decide to have it appear in different locations in future if so desired, possibly even through simple transclusion. That's an easy enough change. -- Eiríkr Útlendi │ Tala við mig 19:10, 23 March 2013 (UTC)

• I'd like to clarify -- my frustration with the current situation is not that due process is somehow getting in the way. Rather, Dan, I wish that you had contacted the editors most involved in this change, i.e. Anatoli, Haplology, and myself, as soon as you noticed that something was afoot that you objected to. Even just in posting to the Grease Pit thread, although you stated your personal reservations, your history as nearly completely uninvolved in Japanese here clearly left Anatoli confused about your chiming in; I know I found myself scratching my head as to why you were getting involved at that point. I wish that you had brought up the issues of due process and the need to allow sufficient time to pass, to allow editors who are not quite so active to see and respond to the proposal before we all moved ahead. Instead, a few more days have lapsed since that point, with more edits made in that time. The thought that this work might have been wasted is a bit frustrating and disappointing.

Anyway, this is not meant as any sort of indictment; rather, I hope to explain my own sense of exasperation. For the record, I think the concerns that Dan brings up are valid and need consideration. -- Eiríkr Útlendi │ Tala við mig 19:51, 23 March 2013 (UTC)

{{rfd}} anyone? Mglovesfun (talk) 19:52, 23 March 2013 (UTC)

You mean we should post a vote on RFDO? That would be, well, unusual... —Μετάknowledge^{discuss/deeds} 20:45, 23 March 2013 (UTC)

Wouldn't be a first though (admittedly, it might be a second). Mglovesfun (talk) 20:46, 23 March 2013 (UTC)

 Done —Μετάknowledge^{discuss/deeds} 00:39, 24 March 2013 (UTC)

We're going to end up having a vote on whether or not it's OK to RFD votes, aren't we? - -sche (discuss) 03:13, 24 March 2013 (UTC)

The discussion sat in Beer Parlour for a long time. User:Dan Polansky failed to take part in it. After an agreement was reached, I have updated Wiktionary:About_Japanese#Romaji_entries with no objections. Haplology, Eirikr and myself started changing entries. Eirikr helped a lot with enhancements in Template:ja-romaji, which is the new basis for romaji entries.

Dan Polansky protested at Wiktionary:Grease_pit/2013/March#Simplification_of_romaji_entries. I gave him a very detailed answer. He failed to respond. Now, he reverted a few dozen of Haplology's edits and demanded us to stop. This looks more like vandalism. He is not going to maintain and synchronise many thousand of romaji, kana and kanji pages but he says "Whether I have been editing Japanese entries is immaterial". I don't think so.

Now, the vote on Mandarin pinyin was modelled on "yánlì" with no definitions. Many pinyin entries don't have definitions, some do. Perhaps pinyin entries should follow the rule more strictly we should remove definitions from entries, which don't follow the rule. The current vote uses " bàndǎotǐ", which has a definition but it's not a model entry and it shouldn't have any definitions. --Anatoli ^{(обсудить}/^вклад) 21:45, 23 March 2013 (UTC)

I believe changes of this scope should only be done via a vote. None of you have explained what is so bad about a vote. You keep on repeating "bureaucracy", as if this were some kind of magic mantra. Romaji entries have been in the state in which they currently are for years; they can wait one more month for a formal evidence of consensus. --Dan Polansky (talk)

Once we have decided on a plan, we like to execute it. A mosquito won't suck out all your blood, but it's still an annoyance. Should we forbear the fly swatter because we didn't have one before? —Μετάknowledge^{discuss/deeds} 00:39, 24 March 2013 (UTC)

He who is short of logic can resort to metaphor and personal attacks. --Dan Polansky (talk) 00:43, 24 March 2013 (UTC)

To be exact, it's a parable. (And calling me "short of logic" is perhaps a personal attack as well. In case you failed to understand my comment, as may well be the case given your offended tone, the "mosquito" represents the vote, not you.) —Μετάknowledge^{discuss/deeds} 00:46, 24 March 2013 (UTC)

The long discussion at BP took a long time to conclude. It lasted from 21 February to 18 March. Sounds like a month to me.

It took so long precisely because we did not agree on the best way to present Japanese to the user, and in particular the user who can't read anything except romaji. We did it for them. I'm sure it would pass a vote. I'm just not to excited about going through the long list of arguments and justifications so that other users could agree to a change that we made for their sake.

I realize that no editors own any of their contributions, but just for some background: a certain amount of the status quo that we have here with romaji entries today exists because I created it. I don't know how many romaji entries I created, but I'd guess around half of them. That experience gives me some perspective on how well those entries work in practice too. Check the histories of Category:Japanese_proper_nouns for some examples. In addition, there were a number of entries that had a ===Romaji=== header, which I converted to its respective part of speech, so the status quo of the time was not 100%. That goes for lots of other areas too, like mixed styles of romanization. Japanese pages were still not entirely consistent in many ways, and there is still work to do.

Just to reiterate Eirkr's point: this change is not destructive. All of the information there should already be on the mirror page (which was just a mirror in the status quo) and in cases where it wasn't, we're merging it with the kana page. The changes show up as a lot of red numbers but it was all (imperfectly and incompletely) duplicated information. Literally copy and paste. That's how I did it with the thousands of pages I made.

Let the record show that I don't want to do a vote. It looks unavoidable now but to be honest I don't like arguments and confrontation so I hope I can avoid it as much as possible until this is settled, or I should say finally settled. --Haplology (talk) 02:01, 24 March 2013 (UTC)

It's not only not destructive, I believe the simplification will help to increase the number of rōmaji entries, similar to the way pinyin simplification did. Rōmaji entries may be created via a bot. Unlike pinyin, which may link to any number of Chinese characters (hanzi), like hiragana, rōmaji entries link to a small number of variants (e.g. hiragana, katakana, seldom mixed script), which in turn, be representations of multiple kanji (e.g. tsuku to follow all the chain). We don't have a miracle program able to keep three types of entries in sync.

Nobody likes confrontation but if it has to be through a vote, then let it be. We need a policy. The opinion of those who actually does the job of working on Japanese entries and has to maintain and follow the policy should be considered first of all. --Anatoli ^{(обсудить}/^вклад) 02:35, 24 March 2013 (UTC)

Examples

Can we see examples linked, or on the page, showing how these headings are usually structured now, and how this proposal would have them?

Is this essentially changing a set of regular entries into “romanized form of” entries?

Does “often level 3 headings” mean sometimes L1 or L2? —Michael Z. 2013-03-23 18:20 z

Old format: tamari. New format: chūkenshu.

Yes.

I have no idea what he was talking about there. —Μετάknowledge^{discuss/deeds} 18:35, 23 March 2013 (UTC)

• @Michael, also see the separate thread above -- all the data that was on romaji entries was also, and will still be, on the corresponding kana entries. This change is basically a move to normalize our data set by avoiding duplication. This eliminates the maintenance problems of ostensibly duplicate entries diverging over time. -- Eiríkr Útlendi │ Tala við mig 19:13, 23 March 2013 (UTC)

The #4 of the current vote contradicts our proposal and itself - #3. Eirikr, Haplology and myself suggest no definitions in romaji, definitions are handled by the kana (hiragana, katakana or mixed script) entry. Compare current "tsuku" (simplified) with "つく", which has all definitions and a previous version of "tsuku" tsuku as of 03:50, 10 March 2011. Having romaji and kana entries get out of sync, let alone the kanji entries ("付く", "着く", etc.). The simplification proposal will eliminate this. --Anatoli ^{(обсудить}/^вклад) 23:26, 23 March 2013 (UTC)

Please edit the vote to propose that which you say you have already achieved consensus for. I have only created a first draft. --Dan Polansky (talk) 00:19, 24 March 2013 (UTC)

Consistency

Why is the format different from the Pinyin and Gothic romanizations?

(And can we please stop saying “soft redirect?” if there is more than one link then this is not analogous to any redirect. The reader has to choose which link to follow, so this is a cross-reference.) —Michael Z. 2013-03-27 18:12 z

1. Different needs for pinyin:

   Pinyin often has a gloss. This is partly because pinyin pages are basically disambig pages, and disambiguation requires letting the user know at least a little bit about each of the various options.

   Even so, if I were an active editor of Chinese romanization entries, I might push for making the pinyin template more like {{ja-romaji}}.

2. About Gothic, I'm not sure. I see that Gothic uses two templates, {{got-rom}} for headword and {{got-romanization of}} for def line. These could similarly be collapsed into a single template that generates both headword and def line. Some editors seem to be quite vehement that these two should be separate and never the twain shall meet; I'm honestly quite puzzled why.
3. (About "soft redirect", I think this nomenclature came about as many [most?] romaji entries only point to a single kana form, and are thus equivalent to redirects. As discussion proceeded, it became clear that some romaji entries actually point to more than one kana form, but by that point, folks were already using the term "soft redirect" to describe how these entries work. ...then again, I notice that at least one other editor describes "alternative form" links as "soft redirects", so I may be mistaken about why this term came to be used.) -- Eiríkr Útlendi │ Tala við mig 20:42, 27 March 2013 (UTC)

Pinyin entries can link to multiple (unlimited number of) hanzi, similar to the way Japanese hiragana can link to multiple kanji. Romaji, on the other hand, can represent a limited number of kana spellings, we decided that 6 parameters was plenty. The look and feel for romaji entries is the same as standard pinyin entries - cf. "yánlì" and "tsuku", only the romaji template makes sure that no other definition is added, just links to kaan (that was the outcome of the pinyin vote as well - no definition other than a link to a hanzi entry), it's easier to add new romaji entries. The term "soft redirect" (rather than a hard redirect, which redirects straight away) is used e.g. for "common misspelling of" entries. By that we mean that the entry displays the kana spelling the romanisation is for, even if it's just one link. We don't have to use the term "soft redirect". --Anatoli ^{(обсудить}/^вклад) 22:27, 27 March 2013 (UTC)

Thank you. (I also asked because I stumbled on a Gothic romanization entry with a POS, but now I can’t find any except af#Gothic. I guess that form is obsolete.) Is there a reason to add “See:” instead of “Romanization of” as in Gothic, or just nothing, as in Pinyin? Maybe better not to introduce a third format, but start a wider discussion about all three. The kana links should probably include the #Japanese fragment, to be consistent.

(To be picky, the colon in romaji entries seems superfluous, and vestigial italic colons in so many pinyin entries look bad.) —Michael Z. 2013-03-27 22:52 z

Personally, I like shorter, and the header already says "Romanization", so I'd prefer it if the Gothic entries also said "See:".

The colon in romaji entries is something I have no strong opinion about. The colon in pinyin entries is there, I think, because a gloss is (or was) supposed to follow. Given the large number of potential entries on pinyin pages, and that they are essentially disambig pages, I'd prefer it if they all had glosses. From a user standpoint, that's much more user-friendly than requiring someone to blindly click on each link until they find what they're looking for. bā, for instance, has 22 entries, and mù has 33; leaving these entirely gloss-less would be a disservice to any users who aren't already at least somewhat fluent in reading Chinese. -- Eiríkr Útlendi │ Tala við mig 23:15, 27 March 2013 (UTC)

(before edit conflict) The entry af#Gothic must be non-standard, see standard Gothic romanisation entries: wisan, wulfs, þuk (examples were given by CodeCat).

The final format of Template:ja-romaji may change (perhaps just minor changes in the wording), it's not up to me but we already have about a thousand entries using it. "Romanization of" may not be a bad thing. Eirikr could probably add the "#Japanese" fragment to the template, can you do it? (I'm bad with templates). Well, not many languages use hiragana and katakana, so it's not a big deal.

As for multiple formats for the vote, I think those who are in favour of the vote in general will not oppose it just because they don't like some minor formatting details. The two people who oppose it outright may have quite a different reason - removing the English translation, PoS, etc. Let's see.

(after edit conflict) Kana entries have short definitions, so there's no loss of info. --Anatoli ^{(обсудить}/^вклад) 23:30, 27 March 2013 (UTC)

• I reworked the template to apply lang codes to the links, and to link specifically to JA entries. Let me know if that suits. -- Eiríkr Útlendi │ Tala við mig 23:59, 27 March 2013 (UTC)

It worked, thank you. --Anatoli ^{(обсудить}/^вклад) 00:39, 28 March 2013 (UTC)

w:Soft redirect is wiki jargon with a significantly different meaning, so I’ve removed it from the proposal where it could only confuse. For those unfamiliar with it, redirect has general-computing meaning that is even more different.

Personally, I like shorter, and the header already says "Romanization",... – we have both longer and shorter prototypes to follow. If we can agree that having a single style is best, then let’s not move forward by adding a third style. I suggest that this proposal should follow either the Pinyin or Gothic format, and we can start a BP discussion about a single format for all three. —Michael Z. 2013-03-28 14:05 z

I can't impose on languages I don't work with and See is shorter than Romanization of. User:CodeCat has developed the templates for Gothic. Gothic would work perfectly with the template similar to {{ja-romaji}}. Pinyin may have too many links, like Japanese hiragana. --Anatoli ^{(обсудить}/^вклад) 22:38, 3 April 2013 (UTC)

Deletion debate

The following discussion has been moved from Wiktionary:Requests 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.

---

Wiktionary:Votes/pl-2013-03/Japanese Romaji romanization - format and content

This vote was created by Dan Polansky (talk • contribs) after consensus had been achieved by all active Japanese editors (and some non-Japanese active editors) in the Beer Parlour to change the format of romaji entries. This vote is excessive and simply adds to the bureaucratic wasteland that we have sought to avoid by having discussions. Several editors have commented at Wiktionary talk:Votes/pl-2013-03/Japanese Romaji romanization - format and content that this vote is unwanted, so I am posting it here per MG's suggestion. —Μετάknowledge^{discuss/deeds} 00:02, 24 March 2013 (UTC)

• I naturally oppose and say keep. Apparent consensus by editors of Japanese entries (whether these editors are Japanese or not) is not enough. If German editors find an agreement to get rid of German compounds, while the native English speakers do not want to get rid of them in this English Wiktionary, the consensus of German editors should not matter. Moreover, changes that involve removal of content from a large set of entries should require a formal evidence of consensus, which a Beer parlour discussion cannot provide. --Dan Polansky (talk) 00:08, 24 March 2013 (UTC)

  I don't fault you for wanting to keep your own vote, but it looks like you don't understand what consensus is. —Μετάknowledge^{discuss/deeds} 00:15, 24 March 2013 (UTC)

  I guess I should not have expected you to actually address any of the points I've raised. --Dan Polansky (talk) 00:27, 24 March 2013 (UTC)

  I hadn't noticed that you'd made any points. Analyzing your paragraph, I see that the first sentence is a declaration of your vote in the RFDO, the next 2 sentences are based on the ridiculous assumption that the discussion was Japanese-editors-only, and your last sentence is an opinion that appears to be related to a misunderstanding of consensus. —Μετάknowledge^{discuss/deeds} 00:36, 24 March 2013 (UTC)

  I do think that voice of the people who work more in a language and have greater knowledge of its intricacies should have more weight in informal consensus building. But I suppose that it probably manifests itself with less experienced editors deferring to more experienced ones. —CodeCat 00:46, 24 March 2013 (UTC)

I don't know if having a vote on the romaji question is necessary or proper. I sympathise with those who think it is not: as it says on Meta, "Polls are evil. Don't vote on everything ... build consensus through polite discussion". As I said last time someone requested deletion of a vote, though: surely holding a RFD-"vote" on the deletion of an active policy vote on a change you support is just taking a shovel and digging deeper into the rabbit hole of strange bureaucracy.

Ruakh once explained to me that the reason even votes in which the options were "support the proposed change" and "retain the status quo" also had "oppose" sections was that people might oppose even having a vote. Recently, I've even seen votes with clearly-titled "oppose this vote" sections. That seems like the thing to do here: add an "oppose having this vote" section, and vote in it rather than the other sections. (Or perhaps it's rabbit-holey to think that's any better than a RFD...) - -sche (discuss) 01:21, 24 March 2013 (UTC)

http://meta.wikimedia.org/wiki/Voting_is_evil is a horrible essay, IMHO. It contains countless dubious claims, none of them supported by evidence. --Dan Polansky (talk) 01:34, 24 March 2013 (UTC)

As a non-Japanese speaker, I just stayed completely out of the discussion as I have nothing useful to add whatsoever. This should not be interpreted as a lack of support from me. Dan, perhaps you could just accept you're a minority of one in this matter and give up now, eh? Mglovesfun (talk) 11:13, 24 March 2013 (UTC)

I agree with Mglovesfun. —Angr 11:31, 24 March 2013 (UTC)

This vote tells me - no matter what you do, you have to check with User:Dan Polansky first! Discussing changes in the Beer Parlour with advanced Japanese editors for a language for about a month (21 Feb. - 18 Mar.) with plenty of chances to give input mean nothing if you don't run it by Dan Polansky. (The {{look}} can be deleted anyway, as far as Dan Polansky is concerned, it failed to attract his attention). The "about" pages like Wiktionary:About Japanese don't matter either, they are not on Dan Polansky's watchlist anyway. Don't worry about his Babel table or his former or future contributions (D.P: "Whether I have been editing Japanese entries is immaterial."), any language is his concern. Even if you give him detailed explanations and reasons, don't assume it's OK to start editing if he doesn't bother to reply. You have to get a specific approval from Dan Polansky, because if you don't, he will mass revert your edits when he chooses without any warning like he did with Haplology's edits.

This vote is nothing but a waste of time and frustration. --Anatoli ^{(обсудить}/^вклад) 22:36, 24 March 2013 (UTC)

I consider this censorship. You guys aren't any better than Hitler was when he was the head of Germany. keep -- Liliana • 21:51, 25 March 2013 (UTC)

w:Godwin's law in action. Still, I think using an rfd on a vote is a bad idea- even one as unpopular as this one. How about we have an rfd on this rfd? Or a duel between a vote to ban rfds on votes and rfd on the vote to ban rfds? Chuck Entz (talk) 22:19, 25 March 2013 (UTC)

That sounds epic! If I create the RFD for that, will you create the vote? XD —Μετάknowledge^{discuss/deeds} 15:56, 26 March 2013 (UTC)

I'll filibuster the motion to proceed, send the undesirable elements to re-education camps, and RFV the vote (because as Liliana knows, you can RFV anything).

It's very bad form to create a vote for anything when there's already a clear consensus (either for it or against it). Such vote creating should result in a nomination here. But, I can't be bothered reading the original discussion so I'm not voting keep or delete here. Mglovesfun (talk) 17:30, 27 March 2013 (UTC)

K – This has brought up some larger questions. It doesn’t really look like consensus to me when some contributors are so upset. Holding the vote shouldn’t interfere too much with the work, and it’s not like we’re on a deadline. —Michael Z. 2013-03-27 18:19 z

I still don't understand the reasons for being so upset when there was plenty of chances to address all issues when the discussion was in Beer Parlour for so long and objections were also answered in GP but there was no reply. It does interfere with work, though, as the most productive and advanced Japanese editors stopped working on Romaji entries waiting for the outcome. Both contributors who strongly object who haven't worked with Japanese entries and don't plan to do it in the future didn't seem to want to know the reasons, only leaving negative remarks. --Anatoli ^{(обсудить}/^вклад) 03:50, 28 March 2013 (UTC)

De-tagged (kept). - -sche (discuss) 18:38, 31 March 2013 (UTC)