If a track has more than 5 artists they must be substituted with Various Artists, similarly if a track is composed of 3 or more individual tracks, the title must be substituded to <Descriptor> Compilation.

The artists of a song must be tracable to existing people.

Any form of vs. such as Vs., VS and the likes are to be written as vs. only.

...and the likes are substituded to and asterisk

Brackets within artist or title fields should be separated from the other text surrounding it, unless there is obvious reason to not do so.

Cyrillic Romanisation: Use BGN/PCGN system for Russian/Cyrillic. Е and е should be romanised as ye if it stands alone or after a, e, ё, и, о, у, ы, э, ю, я, й, ъ, ь. In other cases, it should be romanised as e. ё should be romanised to ye, however, use yo or o to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. For most of the other characters, refer to the first page of this document.

Russian Romanisation: Use BGN/PCGN system for Russian. Е and е should be romanised as ye if it stands alone or after a, e, ё, и, о, у, ы, э, ю, я, й, ъ, ь. In other cases, it should be romanised as e. ё should be romanised to ye, however, use yo or o to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. For most of the other characters, refer to the first page of this document.

Shiguma wrote:

I believe that TV size cuts of songs should have the (TV Size) label on them, regardless of the official source. My reasoning for this is, when you search up a song, having the (TV Size) in the metadata won't affect searching for that song, while also making it very clear that whichever set you are looking at is the short version of a song. If we're bringing common sense into metadata, I don't see why we shouldn't do this.

Pachiru wrote:

"Guest mappers and storyboarders must be added to the tags of a beatmap set." People who makes the hitsounds shouldn't be added to the tags aswell or it's not mendatory?

Nevo wrote:

Pachiru wrote:

"Guest mappers and storyboarders must be added to the tags of a beatmap set." People who makes the hitsounds shouldn't be added to the tags aswell or it's not mendatory?

I definitely feel people who make hitsounds for a set should be required in tags, since they do play a large role in mapsets, also because a lot of hitsounders put forth plenty of time and effort into their work.

Nevo wrote:

Pachiru wrote:

"Guest mappers and storyboarders must be added to the tags of a beatmap set." People who makes the hitsounds shouldn't be added to the tags aswell or it's not mendatory?

I definitely feel people who make hitsounds for a set should be required in tags, since they do play a large role in mapsets, also because a lot of hitsounders put forth plenty of time and effort into their work.

If a track has more than 5 artists they must be substituted with Various Artists, similarly if a track is composed of 3 or more individual tracks, the title must be substituded to <Descriptor> Compilation.

Songs with German umlauts (ü, ö, ä and ß) must be romanised into two-letter combination (ue, oe, ae and ss).

Commas, vs., &, feat./ft., CV: must always use a trailing whitespace. Unless it is a comma, leading whitespace is also required.

If a song or artist are referred to in multiple ways on official sources provided by the artist, the mapper is free to choose any of the romanisations. The only exception to this is if the song already has a mapset in the Ranked Section, in which case the corresponding guideline applies to it.

Only use the Source field if the song comes from, is remixed from or specifically fan-made for a video game, movie, or series. Website names are not an acceptable Artist nor Source.

Guest mappers and storyboarders must be added to the tags of a beatmap set. This is to give credit where credit is due and helping others identify the main contributers of any given beatmap set.

For songs belonging to doujin circles, the circle name must be used over the vocalist or composer, unless these contributors are not part of the circle. In these cases the priority falls on vocalist followed by composer for instrumental songs.

Special unicode characters must be filtered to their nearest standard equivalent or removed from the Romanised Artist and Romanised Title fields within a .osu file. ★ ☆ ⚝ ✩ ✪ ✫ ✬ ✭ 🟉 🟊 ✮ ✯ ✰ and the likes are substituded to an asterisk. Corner Brackets have to be written as quotation marks instead. Other special characters are to be romanised or dropped on case-by-case basis.

ailv wrote:

Only use the Source field if the song comes from, is remixed from or specifically fan-made for a video game, movie, or series. Website names are not an acceptable Artist nor Source.

Does this include if the song was popularized by a specific game/movie/series? I beleive that's how it's being handled rn as well.

Naotoshi wrote:

agre w/ nevo abt hitsounders in tags.

i think forcing tv size is a bit stupid since "cut ver." for example isn't forced if it's not official yet it serves the exact same purpose. it takes players about 3 seconds to see if a map is tv size or not, and in many cases dropping it makes the title look much cleaner anyway.

Sieg wrote:

alright

Cyrillic Romanisation: Use BGN/PCGN system for Russian/Cyrillic. Е and е should be romanised as ye if it stands alone or after a, e, ё, и, о, у, ы, э, ю, я, й, ъ, ь. In other cases, it should be romanised as e. ё should be romanised to ye, however, use yo or o to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. For most of the other characters, refer to the first page of this document.

I suggest to remove "Cyrillic" and leave it to case by case since current wording is ambiguous trying to cover all Cyrillic languages with different phonetics and different scenarios of romanisation, which is obviously wrong approach. So, reword it to

Russian Romanisation: Use BGN/PCGN system for Russian. Е and е should be romanised as ye if it stands alone or after a, e, ё, и, о, у, ы, э, ю, я, й, ъ, ь. In other cases, it should be romanised as e. ё should be romanised to ye, however, use yo or o to avoid usage of special characters. Ignore any other rules in the file provided, these are either irrelevant or wouldn't help in the game. If an artist uses a preferred romanisation, follow it regardless of this rule. For most of the other characters, refer to the first page of this document.

J1NX1337 wrote:

Brackets within artist or title fields should be separated from the other text surrounding it, unless there is obvious reason to not do so.

-> unless there is an obvious reason not to do so.
Slight grammar correction.

Kurai wrote:

- Cyrillic Romanisation should follow the BGN/PCGN system (except for the letter ё in Russian which should follow the GOST 2002(B) system). Read more here: http://up.kuraip.net/032209ex3724.pdf

Pachiru wrote:

For tags, from draft: "This is to give credit where credit is due and helping others identify the main contributers of any given beatmap set."

Since we should give feedback to the main contributors of the set, we need to put the person who made the hitsounding, which is an important part of the map, because without hitsounds the map is unrankable. If we give credit to someone who made something optional like Storyboard, we should add the hitsounds person to the tag aswell.

(also, don't forget to fix the typo on: "contributers" → "contributors")

ailv wrote:

If a track has more than 5 artists they must be substituted with Various Artists, similarly if a track is composed of 3 or more individual tracks, the title must be substituded to <Descriptor> Compilation.

What about albums? E.g If I were to map https://soundcloud.com/owslaofficial/se ... nctuary-ep the entirety of this ep mixed into one track, I can't see why I'd have to name it "Sanctuary EP Compilation" over just "Sanctuary EP".
suggest an alternative, this is a fair point

Songs with German umlauts (ü, ö, ä and ß) must be romanised into two-letter combination (ue, oe, ae and ss).

This doesn't make sense in certain cases, for example https://osu.ppy.sh/s/723626 wouldn't make sense as the "Ü" is simply a stylizing choice, it's intended to be a U with some fancy shmancy stuff.
limiting it to romanisation of german then because german words are affected

Commas, vs., &, feat./ft., CV: must always use a trailing whitespace. Unless it is a comma, leading whitespace is also required.

does this include "ft" or "feat" without a "."
now it does

If a song or artist are referred to in multiple ways on official sources provided by the artist, the mapper is free to choose any of the romanisations. The only exception to this is if the song already has a mapset in the Ranked Section, in which case the corresponding guideline applies to it.

How does this apply to https://osu.ppy.sh/p/beatmaplist?q=diao%20ye%20zong which has had maps ranked under "Diao Ye Zong" and "RD-Sounds".
if the song youre mapping has 2 ranked sets with both then you can go back to choosing, otherwise follow ranked unless that one was wrong

Only use the Source field if the song comes from, is remixed from or specifically fan-made for a video game, movie, or series. Website names are not an acceptable Artist nor Source.

Does this include if the song was popularized by a specific game/movie/series? I beleive that's how it's being handled rn as well.
example needed?

Guest mappers and storyboarders must be added to the tags of a beatmap set. This is to give credit where credit is due and helping others identify the main contributers of any given beatmap set.

Does this need to be extended to previous usernames?
does it say so? no.

For songs belonging to doujin circles, the circle name must be used over the vocalist or composer, unless these contributors are not part of the circle. In these cases the priority falls on vocalist followed by composer for instrumental songs.

I'm not sure if I'm interpreting this wrong, but shouldn't it be more something like guest composer + circle, to give accurate credit in those cases?
usually when that happens the song is a guest on an album anyways but im not sure myself more debate required.

Special unicode characters must be filtered to their nearest standard equivalent or removed from the Romanised Artist and Romanised Title fields within a .osu file. ★ ☆ ⚝ ✩ ✪ ✫ ✬ ✭ 🟉 🟊 ✮ ✯ ✰ and the likes are substituded to an asterisk. Corner Brackets have to be written as quotation marks instead. Other special characters are to be romanised or dropped on case-by-case basis.

I believe there should be some clarification between ' ' and " ", being used for romanization of 「 」, see https://osu.ppy.sh/s/735097, https://osu.ppy.sh/s/538136. Since British and American ways differ.
then clarify instead of saying that?

The same applies to the Source field if a romanised Source is preferred by the mapper.

ailv wrote:

If a track has more than 5 artists they must be substituted with Various Artists, similarly if a track is composed of 3 or more individual tracks, the title must be substituded to <Descriptor> Compilation.

What about albums? E.g If I were to map https://soundcloud.com/owslaofficial/se ... nctuary-ep the entirety of this ep mixed into one track, I can't see why I'd have to name it "Sanctuary EP Compilation" over just "Sanctuary EP".
suggest an alternative, this is a fair point

Compilation should be used in cases where the songs are not already part of a organized set of songs.

Songs with German umlauts (ü, ö, ä and ß) must be romanised into two-letter combination (ue, oe, ae and ss).

This doesn't make sense in certain cases, for example https://osu.ppy.sh/s/723626 wouldn't make sense as the "Ü" is simply a stylizing choice, it's intended to be a U with some fancy shmancy stuff.
limiting it to romanisation of german then because german words are affected

Can we state that this appleis speficially to Songs with a German title or artist fields then? Currently the wording implies that any song with those german umlaut characters would need to be romanized as such.

Guest mappers and storyboarders must be added to the tags of a beatmap set. This is to give credit where credit is due and helping others identify the main contributers of any given beatmap set.

Does this need to be extended to previous usernames?
does it say so? no.

I've seen discussion about this before, I think I've seen cases where the previous username was added to the online tags so a DQ wouldn't be needed. I think further discussion would be useful.

Special unicode characters must be filtered to their nearest standard equivalent or removed from the Romanised Artist and Romanised Title fields within a .osu file. ★ ☆ ⚝ ✩ ✪ ✫ ✬ ✭ 🟉 🟊 ✮ ✯ ✰ and the likes are substituded to an asterisk. Corner Brackets have to be written as quotation marks instead. Other special characters are to be romanised or dropped on case-by-case basis.

I believe there should be some clarification between ' ' and " ", being used for romanization of 「 」, see https://osu.ppy.sh/s/735097, https://osu.ppy.sh/s/538136. Since British and American ways differ.
then clarify instead of saying that? I don't know which one is better, i'd propose that bot are acceptable and up to choice.

the thing part nine wrote:

Word-by-word Romanisation: Each character must be romanised into a single, capitalised, separated word. Refer to this thread for examples and supplementary information.

the thing part thirty wrote:

Guest mappers, storyboarders, and hitsounders must be added to the tags of a beatmap set. This is to give credit where credit is due and helping others identify the main contributors of any given beatmap set.

the thing part forty two wrote:

Commas, vs., &, any variations of feat./ft., CV: must always use a trailing whitespace. Unless it is a comma, leading whitespace is also required.

Okoratu wrote:

4 yea idk either

If the creator of the mapset has done major edits to the .mp3, they are free to name it appropriately to signal that this song is a special version. In this case the original songs must still be clearly indicated in order for players to be able to search for the original songs.

Corner Brackets have to be written as quotation marks instead.

Other special characters are to be romanised or dropped on case-by-case basis.

Brackets within artist or title fields should be separated from the other text surrounding it, unless there is obvious reason not to do so. Reasoning like this would include syntactical use of brackets and the general typesetting of a song title or artist using them without whitespaces often and consistently across multiple platforms.

ailv wrote:

Corner Brackets have to be written as quotation marks instead.

Add an example of what a corner bracket is "⸤ ⸥" and " ｢ ｣ " since there are multiple forms.

Noffy wrote:

That's why it's in the glossary

Okoratu wrote:

also the source is a unicode field as such it can hold anything we want it to -> the mapper should have the choice to decide which one is shown in the client

Okoratu wrote:

Kurai wrote:

- Cyrillic Romanisation should follow the BGN/PCGN system (except for the letter ё in Russian which should follow the GOST 2002(B) system). Read more here: http://up.kuraip.net/032209ex3724.pdf

can some russians say something about http://up.kuraip.net/032209ex3724.pdf ? it seems to make sense and encompass all things said

pdf wrote:

Hopefully, the BGN/PCGN systems have been built so that Cyrillic can be rendered by using only the basic letters and punctuation found on English language keyboards.

CrystilonZ wrote:

just as third party stuff arent accepted as refs imo crunchyroll shouldn't be accepted as well i guess

for cases like games that are released in a lot of regions english names should be okay
ea. both ポケモン超不思議のダンジョン and Pokémon Super Mystery Dungeon are fine

Crunchyroll wrote:

officially-licensed content from leading Asian media producers directly to viewers translated professionally in multiple languages

Proposed Rules wrote:

1. Languages in Chinese language family must be Romanised accordingly. Do not Romanise Cantonese texts with Hanyu Pinyin method of Romanisation.
2. Songs with Mandarin titles and/or Mandarin artists must use the Hanyu Pinyin method of Romanisation when there is no Romanisation or translation information listed by an official source. The ü vowel should be Romanised into u and all diacritical tone marks should be omitted because of the technical limitations resulting from the limited amount of characters allowed in the Romanised title/artist fields.
3. For capitalisation and word separation, refer to The Basic Rules of the Chinese Phonetic Alphabet Orthography (汉语拼音正词法基本规则/漢語拼音正詞法基本規則). In short, generally every word should be separated and capitalised. Surname and first name are separated using a space and are capitalised.
4. Particles (助词) are written separately and should not be capitalised.

Proposed Rules wrote:

1. Songs with Mandarin titles and/or Mandarin artists must use the Hanyu Pinyin method of Romanisation when there is no Romanisation or translation information listed by an official source. The ü vowel should be Romanised into u and all diacritical tone marks should be omitted because of the technical limitations resulting from the limited amount of characters allowed in the Romanised title/artist fields. There is a vowel that is "u", using "u" for "ü" just mess up them, currently there are not better choice other than "v". And only at very a few cases, it may use 'yu' for people name, and it's arguable, but I am not able to consider this is suitable to represent all the "ü"

Wafu wrote:

1. Current system
   
   1. Titles are easy read ✘ (most of people will read every syllable as if it was one word)
   2. Titles are easy to remember ✘ (words are easier to remember than separate syllables, humans remember the words easier by their shape)
   3. Fits the rules of Latin script (Romanisation = writing words from other script to Latin/Roman script) ✘ (Latin script is alphabetical, therefore separating each syllable doesn't make sense and doesn't read well for majority of Latin script)
   4. Fits the rules of Chinese script ✘ (Impossible, if you want to make it "fit" to the Chinese script, you would have to replace each character with one logogram, Latin alphabet doesn't have logograms. Chinese is also not syllabary script, so separating each syllable again doesn't make sense.)
   5. Differentiates between different Romanisations and meanings of the same sequences of characters. ✘
   6. Includes tones in Romanised text ✘ (Impossible with characters which we are limited to. You could use "a1", "a2" (redundant) etc., but that would make the text incomprehensible, majority of people wouldn't even know how to pronounce it)
   7. Doesn't replace characters with others which have no evidence of being similar to the intended character. ✘ (ü is replaced with v, which doesn't seem to be supported by any logical argument)
   8. You can use a different Romanisation system for dialects where the current system wouldn't work at all ✘
   9. Isn't related to politics ✘ (Impossible, picking any Romanisation system is picking a side, every Romanisation system is related to politics)
2. Proposal
   
   1. Titles are easy read ✔ sorry but as a Chinese speaker I don't think they are not easier to read than current seperated romanisation format
   2. Titles are easy to remember ✔ They are not easier to remember than current seperated romanisation from a Chinese speaker side as well
   3. Fits the rules of Latin script (Romanisation = writing words from other script to Latin/Roman script) ✔Latin script isn't fittable to Chinese, Pinyin system is much better from my side too
   4. Fits the rules of Chinese script ✘ (Impossible, if you want to make it "fit" to the Chinese script, you would have to replace each character with one logogram, Latin alphabet doesn't have logograms.)
   5. Differentiates between different Romanisations and meanings of the same sequences of characters. ✔explained above, for Chinese characters, Pinyin or any other romanisation system doesn't have any meaning at all, it's just a way to use as mark for Mandarin
   6. Includes tones in Romanised text ✘ (Impossible with characters which we are limited to. You could use "a1", "a2" (redundant) etc., but that would make the text incomprehensible, majority of people wouldn't even know how to pronounce it)
   7. Doesn't replace characters with others based on no evidence that they are similar to the intended character. ✔ as the pinyin or anyother romanisation/latin letters (they are just used as mark as I said)doesn't stand for meaning, this feels unnecessary
   8. You can use a different Romanisation system for dialects where the current system wouldn't work at all ✔ Dialects don't have an official way of writing formally, even in HK, schools teach the Mandarin grammar and write standard Chinese grammar while only use cantonese as a pronuciation. There wouldn't be songs that use dialects as song titile and artist, so they don't need to be romanised at any case.
   9. Isn't related to politics ✘ (Impossible, picking any Romanisation system is picking a side, every Romanisation system is related to politics)

Fycho wrote:

Below are that I disagree: (blue are my replies)

Wafu wrote:

Proposal

1. Titles are easy read ✔ sorry but as a Chinese speaker I don't think they are not easier to read than current seperated romanisation form As stated above they are easier for the majority of the player base and they are not harder to read for the Chinese. That means no cons just pros.
2. Titles are easy to remember ✔ They are not easier to remember than current seperated romanisation from a Chinese Speaker side as well
3. Fits the rules of Latin script (Romanisation = writing words from other script to Latin/Roman script) ✔Latin script isn't fittable to Chinese, Pinyin system is much better from my side too
4. Fits the rules of Chinese script ✘ (Impossible, if you want to make it "fit" to the Chinese script, you would have to replace each character with one logogram, Latin alphabet doesn't have logograms.)
5. Differentiates between different Romanisations and meanings of the same sequences of characters. ✔explained above, for Chinese characters, Pinyin or any other romanisation system doesn't have any meaning at all, it's just a way to use as mark for Mandarin I think the point here is about homophones and stuff. Same sequence of pronunciation might produce two (or more) different meanings. Separating Romanised titles into words should help with the comprehensibility.
6. Includes tones in Romanised text ✘ (Impossible with characters which we are limited to. You could use "a1", "a2" (redundant) etc., but that would make the text incomprehensible, majority of people wouldn't even know how to pronounce it)
7. Doesn't replace characters with others based on no evidence that they are similar to the intended character. ✔ as the pinyin or anyother romanisation/latin letters (they are just used as mark as I said)doesn't stand for meaning, this feels unnecessary speaking about u and v here. v is just impossible to pronounce. I'm always open for a better alternative.
8. You can use a different Romanisation system for dialects where the current system wouldn't work at all ✔ This is only about Mandarin, dialects are not included. (Cantonese, Wu-Chinese(Shanghainese, Suzhou Hua, Wuxi , HangZhou), Jiang–Huai Mandarin, Southern Fujian Dialect, Hakka Dialect, etc don't need to be discussed for now this is speaking about the new rule. Languages in Chinese language family must be Romanised accordingly. This opens room for other Chinese languages to use suitable Romanisation systems, not restricting all Chinese languages to one bad system which may or may not fit the language.
9. Isn't related to politics ✘ (Impossible, picking any Romanisation system is picking a side, every Romanisation system is related to politics)

CrystilonZ wrote:

I think the point here is about homophones and stuff. Same sequence of pronunciation might produce two (or more) different meanings. Separating Romanised titles into words should help with the comprehensibility.

CrystilonZ wrote:

Other languages that use the Chinese script are irrelevant to this proposal.
We are only talking about Standard Mandarin here and Mandarin is not equivalent to Chinese.
We only use 'Chinese' in the draft for simplicity. The wording will be changed if this is implemented.

1. is this a Mandarin Chinese song?
   - maybe from official settings or sites, not a big deal. but will not do if you map some cult song.
2. how to get the right romanized characters?
   - ask some Chinese staff/mapper/player? i doult any of them have time/ability to do it.
3. how to make sure those things i got is correct?
   - some kind of same as the one above, if that person exsist and can do his job endlessly, he will be really welcomed to this system.

Proposed Rules wrote:

The ü vowel should be Romanised into u and all diacritical tone marks should be omitted because of the technical limitations resulting from the limited amount of characters allowed in the Romanised title/artist fields.

CrystilonZ wrote:

speaking about u and v here. v is just impossible to pronounce. I'm always open for a better alternative.

Previous Discussion wrote:

1. Current system
   
   1. Titles are easy read ✘ (most of people will read every syllable as if it was one word)
   2. Titles are easy to remember ✘ (words are easier to remember than separate syllables, humans remember the words easier by their shape)
2. Proposal
   
   1. Titles are easy read ✔
   2. Titles are easy to remember ✔

Previous Discussion wrote:

For my thoughts on the matter I don't understand how romanising 学不会 to Xue Bu Hui is more informative than Xue Buhui or Xuebuhui. Word separation can be ambiguous at times but whether you write Xue Buhui or Xuebuhui it's more informative than the Xue Bu Hui according to the current RC.

Previous Discussion wrote:

1. Fits the rules of Chinese script ✘ (Impossible, if you want to make it "fit" to the Chinese script, you would have to replace each character with one logogram, Latin alphabet doesn't have logograms.)
2. Differentiates between different Romanisations and meanings of the same sequences of characters. ✔
3. Doesn't replace characters with others based on no evidence that they are similar to the intended character. ✔

1. Proposal
   
   1. Titles are easy read ✘
   2. Titles are easy to remember ✘
   3. Fits the rules of Chinese script ✘
   4. Differentiates between different Romanisations and meanings of the same sequences of characters. ✘
   5. Includes tones in Romanised text ✘
   6. Doesn't replace characters with others based on no evidence that they are similar to the intended character. ✘
   7. Isn't related to politics ✘
   8. Easy for non-Mandarin speakers/players to search in the beatmap list. ✘
   9. Avoid mislead players about the meaning of title. ✘
   10. Provide a practical or feasible method of romanization. ✘

Hollow Wings wrote:

1. In automatic romanizing working progress, there're two ways for Chinese Romanisation:
a. semi-automatic romanisation from Chinese words separated by following proper rules.
b. automatic romanisation from Chinese characters one by one.
2. During this period of time, most of other countries aside of PRC can't fully accept that romanizing Chinese characters into separated words according to combinations between Chinese characters, because the works of finding and dealing with the concept of Chinese words are complex, also the grammar of Chinese sentence can even blur it.
after thousand of thoughts, they decide to do the romanization work from Chinese characters one by one.

Hollow Wings wrote:

1. Egyptian hieroglyphs (eg. Ancient Egyptian) ←already dead
2. Cuneiform script (eg. Ancient Sumerian) ←already dead
3. Seal hieroglyphs (eg. Ancient Indian) ←already dead
4. Maya hieroglyphs (eg. Ancient Mayan) ←already dead
5. Chinese characters (eg. Chinese)

and NO MORE.
if you want to know why language system is like that, then that's a long story, i wont start telling them here.
the reason i pick up those truth above, is because i want you guys know the chinese language's specificity and leading to how different romanisation is done between alphabetic language and ideographic language.

Hollow Wings wrote:

however, this is not reversible.

abraker wrote:

Any thoughts about mapping style or patterns the maps have being in tags?

1. If we romanise Chinese title in word-by-word way(each character must be romanised into a single, capitalised, separated word) or generally every word should be separated and capitalised according to The Basic Rules of the Chinese Phonetic Alphabet Orthography.
2. If using "yu" or "u" for the romanisation of the vowel "ü".
3. If we need to distinguish dialects from Mandarin in romanisation.