@42nevin said:

@lineker said: I think in your case, there are easy technical solutions where a fallback option (zh-SG) would be in play in all of these cases. I think that is the best way forward for you, even though it requires a little bit of work with contacting whatever app you are using.

@sdjk said: There are only options such as Simplified Chinese, Traditional Chinese, and Hong Kong Chinese. Today's streaming media platforms have a large amount of data available in multiple languages. In different regions, as long as different user languages are selected, the corresponding official translation titles will appear, which has long weakened the concept of local release.

I'm glad that such discussion is finally heading towards the right direction. The core of the problem is that simplified Chinese is not represented in TMDb, but all four options are generically called “Chinese”. Ideally there should be a system that differentiates script and region subtags of a language: zh-Hans/-Hant first and then -CN/-HK/-SG/-TW:

1. Simplfied Chinese (zh-Hans) is used officially in mainland China (CN), Singapore (SG) and unofficially Malaysia;
2. Traditional Chinese (zh-Hant) is used officially in Hong Kong (HK), Taiwan (TW) and Macau;
3. Simplified Chinese can be understood and sometimes used in areas where traditional Chinese is the official language and vice versa;
4. zh-HK could mean zh-cmn-HK or zh-yue-HK. Yue/Cantonese will have different Romanization and sometimes different written form than Mandarin;
5. Because many movies/shows will only be released in or outside of mainland China, and because TMDb prioritises official release, users may want zh-Hant-CN, or zh-Hans-TW, or zh-Hans-HK as an option.

When it comes to Chinese releases and Chinese users, the primary consideration is not of region, but whether it's Mandarin or Yue/Cantonese, simplified or traditional. The logic that applies to nation state in a typical Western sense is not entirely applicable here. If Hans or Hant is sufficient enough to differentiate a title, then there is no point in CN/HK/SG/TW. These are only needed if furthur distinction is needed (say, if the movie/show has different titles in Hong Kong and Taiwan)—for textual information anyway, as opposed to release date, content rating, etc that is associated with regions. Since translation is by definition the process or instance concerning a language pair, and languages exist and are used before any tagging system, then it can only mean one thing if an existing translation is rejected by the rule of that system: something is wrong during the tagging phase. Either it was too prescriptive or obsolete to be useful among the real language users or said users could not understand the tags [1].

@lineker said: OK, so you want us to remove Singapore and create one union with China and Singapore so the same type of rule can be used for simplified Chinese?

I’ve rarely seen difference in titles used in mainland China and Singapore though, so it might not be a completely bad idea to merge them after all (by the same token of en-US covering all English countries). Plex, for example, only has Chinese, Chinese (Hong Kong) and Chinese (Taiwan), not an accurate nomenclature but it works better in real world usage. The few that are different can be accommodated by alternative title depending on release date.

For those interested in this topic, may I suggest reading W3C's Understanding the New Language Tags and this blog post titled "The Globalism of Apple TV+"? The latter is in Chinese, so I took the liberty to edit and translate the parts relevant to our discussion in footnotes [2].

@lineker said: Language codes and country matches used on TMDb follow ISO 693 which has some limitations.

I believe you meant ISO 639? This may sound pedantic, but "en-US" or "zh-CN" is a combination tag. ISO 639 is about language only, not country/region, which falls under ISO 3166, and I'm suggesting incorporating ISO 15924 (script). This would, in addition, solve many other problems such as how Romanised title can only be added to Alternatiive Titles without more rigorous control. Russian, Japanese… for example can all benefit from this. By the way, the code for Cantonese used by TMDb is cn-CN, which does comply with ISO 639 (should be yue or zh-yue).

@lineker said: However, that would require a change of the rules, saying that, for example, a Swedish translated title doesn't have to be in use in Sweden. That would allow users to add unofficial translated titles to a lot of movies. I'm not sure that would be a good idea in the long run, even though users would be happy in the short perspective.

It's not fair to compare this to, say, a Swedish translated title, because Swedish is the official language of only Sweden and part of Finland, so to have an official translation would imply that it's used in Sweden (the nation state logic), at least before the streaming age. Plus, there's no scriptual variation involved (unlike Norwegian). As @sdjk has pointed out, it's much more comparable to English used in US, UK…India, etc. Although, I am aware of the fact that there is currently only one locale (en-US) for all variants of English but four for Chinese, so the same justification for English title to be used as long as it's released in any one English-speaking country may not apply, but we don't have a region-agnostic language option which is rather desirable in these kind of situations. Nor do we have zh-MO (Macau) while we are at it. The separation is, not to put too fine a point on it, half-knowledge, and without proper marking, almost insensitive. The tag "zh-CN", as has been noted, encompases ambiguity in the first place. Does that mean Chinese used only in mainland China (if so, which dialect? which script? classical or modern?) or just simplified Chinese [3], a legacy workaround substitute for zh-Hans? In many other websites and softwares, it's often the latter (and usually no zh-SG, making it the de facto synonym), but it's obviously not the case for TMDb, hence the complaints. Both definitions are imperfect though. One cannot impose an artificial classification [4] on a natural (macro)language, binding it to a specific region, and then equate it with that also used in other regions. On the other hand, there’s also technically no such thing (or not a single entity) as “Chinese used only in mainland China”. The title used in mainland China is not the title in Chinese (mainland China) [5], and that's not even considering the rabbit hole of the word China. Yes, I understand from a LIS point of view, that such an artificial classification is often needed, but a normal user is unlikely to see things this way. Because of the ambiguity and perhaps failure of communication, Chinese users are angered not only as it's extremely inconvenient, but it also feels an attack to or denial of them as a civilisation. Either one makes an ad hoc solution for "zh-CN" or one stops using the tag "zh-CN" all together (or stops calling it Chinese/汉语 [6], but use the region name).

Despite TMDb's best intent, it's still, I'm afraid, very much English-centric, and thus fails to deal with some specific requirements arised from the complicated nature of multi-languages and cultures. I've argued elsewhere that it's more reasonable and convenient to leave the translated title blank for users whose languages are of the same writing system as the original title. French users might not understand the English title but can still recognise and maybe pronounce it. And most apps seem to just work fine with any language in Latin alphabet. Another example would be how Chinese users, even if they can't speak Japanese, usually prefer to have Japanese songs in its original script rather than romanised or even translated.

Most of us also prefer the original titles over the (often bad!) official translated titles. But when you are editing, please try to put your personal preferences aside. (Bible)

But it's not the case for a distinctively different writing system.

Exception: a romanized title can be used as a temporary translated En-US title for entries with a non-Roman original title (e.g. a Japanese, Malayalam or Ukranian movie). (Bible)

@banana_girl said: For non-English content without an official release, you can use semi-official titles like an English festival title or an English title use[d] by the producer/foreign distributor.

These suggest that TMDb acknowledges this issue, but only has remedies for English (United States). Why should English speakers be allowed the extra privilege of literacy when people from other writing systems has also long practiced similar approach? Shouldn't the same principle be applied reciprocally?

To make the matter worse, there are geopolitical factors forbidding such movies/shows that do offer official translation from a proper release inside certain regions, which is all the more reason to warrant a zh-Hans tag instead of or as fallback of zh-CN. Indeed, this is not TMDb's responsibility and language fallback is more of a client-side problem, but it would be better for everyone if a solution can be reached at the upstream level, perhaps having the title in zh-CN fallback to zh-SG similar to how the title of a TV episode fallbacks to "Episode X" if there's no input—that is, if expanding language tags to zh-Hans is not viable. Another solution would be to allow alternative title to be displayed/fetched (with proper mark of course) if there's no "official" one.

For reference, check out Wikipedia's solution for the similar problem. An automatic character conversion may be needed if there's no official Singaporean release either, in which case Chinese (simplified) should fallback to text in Chinese (traditional, Taiwan/Hong Kong) but properply converted, and Chinese (traditional), vice versa (having text converted from simplified Chinese if something is only released in it). I'm afraid in the long run that a multi-level fallback/conversion system at TMDb's end is the only way to ensure database integrity and user experience for all demographics of Chinese.

To summerise all my proposed solutions in order of desirability:

1. Use a system of appropriate language tags (zh-Hans, zh-Hant, etc.) with multi-level fallback and conversion.
2. Having zh-CN and zh-SG fallback to each other, or make alternative tilte more prominent in the absence of translated title. (Remember how exception was made to accomodate English users who may have trouble reading and typing non-Latin characters?)
3. Merge zh-CN and zh-SG into zh-Hans.
4. Continue the current practice, but rename each locale to its respective region name.

---

Footnotes

• [1]: Not many people can navigate a library and find what they want with only Dewey decimals.
• [2]: "Most of the movies and shows on Apple TV+ are translated to dozens of languages including captions and dubs on launch, which is unusual in the history of media. Hollywood movies with big budgets might get the same treatment in preparation for releases across different countries, but TV networks and even Netflix would only provide at most 3–5 languages specific to a regional audience. More importantly in the case of Apple TV+, said translations are available globally, no matter where you are. You can, for example, watch Ted Lasso in Telugu even if you are not remotely in India (or South Africa)."
• [3]: As far as I know, there used to only be three codes in TMDb: zh-CN, zh-HK, zh-TW; I don't know why zh-SG was not there at first or why it was added but you can see why people would assume zh-CN means simplified Chinese, just as about everywhere else.
• [4]: https://ib.bioninja.com.au/standard-level/topic-5-evolution-and-biodi/53-classification-of-biodiv/classification.html
• [5]: Take this movie as an example, the title used officially in mainland China is 恋爱大赢家 or 戀愛大贏家, same as that in Hong Kong, and the title in Chinese (mainland China) is 恋爱大赢家 or 喜欢你喜欢我. The title used in Taiwan is 喜歡你喜歡我 or 喜欢你喜欢我. Sure, this is being nitpicking, and in reality, we have all sorts of assumption in place to avoid this kind of conundrum. But assumption can be dangerous if different parties have different ones. An exact string can only be identified by multiple conditions such as "a title used in mainland China in simplified Chinese". That of course won't fit into most interfaces, but "Chinese (zh-CN)" seems to be too concise.
• [6]: One shouldn't use 汉语/漢語 anyway when referring to the written text of Chinese. The accurate designation would be 中文. Simplified Chinese is 简体中文, but never 简体汉语.

你的英文真好，理由充分，数据翔实，但是再一看日期：「于 2022 年 06 月 14 日 7:59上午」，所以tmdb的管理到底是坚持规则呢，还是别有用心呢，不言而喻。

我就纳闷了，怎么翻译还有国别了，难道不是因为有了翻译，语言才有了交流，到了TMD这边就开始什么大陆没有上映，虽然这点问题对我影响不太大，但是我就是要说你们TMD

你的英文真好，理由充分，数据翔实，但是再一看日期：「于 2022 年 06 月 14 日 7:59上午」，所以tmdb的管理到底是坚持规则呢，还是别有用心呢，不言而喻。

zzzq