Add a meta tag for language and let everyone enjoy the public timelines

[Lomplac]

Recently many french speakers came on mastodon.social
Then the public timelines became multilingual, which is a problem for most people who only want to read understandable toots

A way to fix that problem would be to prohibit other languages than english on the public timelines, but i don't think it's fair for non-english speakers who also want to meet other people and enjoy those timelines
What i suggest is to add a new meta tag which specify the language of every toot, and the possibility of filtering the public timelines

We could determine which language a toot contains automatically (with a bot or something), or just let the users choose in the settings in which language they toot
For multilingual users a button as below would be great

April 14th, update that includes some suggestions:

• ISO 639-3 acronyms
• Language option in the toot box
• Option in settings to pick the default languages there will be in the toot box
• Option in the settings that will indicate the languages of every toot
• New notification-like filter but for langages
• Temporary and permanent language filters
• Filter that allow to show only specific languages
• Shortcut to mute the language of a toot
• No filter by default
• Toots without text won’t be filtred

[kennethlarsen]

This would be amazing and I think the key here is automatic language detection since most people would forget or be to lazy to assign it manually. This would be possible with libraries like:

https://github.com/wooorm/franc
or
https://github.com/peterc/whatlanguage

[kennethlarsen]

This would be amazing and I think the key here is automatic language detection since most people would forget or be to lazy to assign it manually. This would be possible with libraries like:

https://github.com/wooorm/franc
or
https://github.com/peterc/whatlanguage

[zuglufttier]

Please, no automatic detection! The browser will send a default language and 95% of people will be happy with that.
I, and I'm surely not the only one, write in two languages and want to keep this ability, just add the option to change the default language which could be set by the browser.

[zuglufttier]

Please, no automatic detection! The browser will send a default language and 95% of people will be happy with that.
I, and I'm surely not the only one, write in two languages and want to keep this ability, just add the option to change the default language which could be set by the browser.

[kennethlarsen]

I agree that using the default browser language will not work. This might be the way to do it:

• Add an option to manually assign language in the toot box
• Add option in settings to pick default language in the toot box
• Sort streams by language

This can then later be refactored into automatic detection by using one of libraries mentioned before. This does not mean detecting the default browser language but instead analyzing the toot before it's being sent and then setting the language meta tag based on the specific toot. This way you can write toots in all the languages you want to and still have automatic detection.

[kennethlarsen]

I agree that using the default browser language will not work. This might be the way to do it:

• Add an option to manually assign language in the toot box
• Add option in settings to pick default language in the toot box
• Sort streams by language

This can then later be refactored into automatic detection by using one of libraries mentioned before. This does not mean detecting the default browser language but instead analyzing the toot before it's being sent and then setting the language meta tag based on the specific toot. This way you can write toots in all the languages you want to and still have automatic detection.

[Ravenbird]

i also prefer that the user can choose the language of the post. But we need a solution for this problem.

[Ravenbird]

i also prefer that the user can choose the language of the post. But we need a solution for this problem.

[nclm]

All of this is a very good idea.

An easy way to this for the user without adding specific filters would be to build a new setting in the options panel: “Languages I understand”. The user sets there the languages they can read and/or write.

The write box would only list languages from these, and if set, the timelines would be filtered to display only toots in these (without segregating them), with an option to “Display all languages” in the columns filters.

If not set yet, the write box would default to the browser’s language and/or automatic detection, with a visible invitation to set the setting, and the timelines would not be filtered at all.

Alternatively for new users, it could be something asked at sign-up.

[nclm]

All of this is a very good idea.

An easy way to this for the user without adding specific filters would be to build a new setting in the options panel: “Languages I understand”. The user sets there the languages they can read and/or write.

The write box would only list languages from these, and if set, the timelines would be filtered to display only toots in these (without segregating them), with an option to “Display all languages” in the columns filters.

If not set yet, the write box would default to the browser’s language and/or automatic detection, with a visible invitation to set the setting, and the timelines would not be filtered at all.

Alternatively for new users, it could be something asked at sign-up.

[KonomiKitten]

I would really like to see this idea happen at the moment I just mute people who I can't read the languages of which is rather tedious, having the ability to select which languages I can understand in my profile and see only those would be beneficial for everyone.

[KonomiKitten]

I would really like to see this idea happen at the moment I just mute people who I can't read the languages of which is rather tedious, having the ability to select which languages I can understand in my profile and see only those would be beneficial for everyone.

[jeanbaptisteb]

If this feature is implemented (which would be great), it may be relevant to use the ISO 639-3 code list (https://en.wikipedia.org/wiki/ISO_639-3).

This standard includes the possibility to tag a language as "undetermined", or to tag the content as not including text. Think for example about people just tooting a photograph (or a more complicated case: someone tooting just a link to a video with subtitles in multiple languages).

ISO 639-3 also includes codes for some artificial, rare, or ancient languages. Contrary to ISO 639-1 (for example), this standard tries to be as comprehensive as possible; it would be more inclusive to use it.

Besides the question of the standard to choose, another thing is to do is to separate the notions of "languages I'm able to understand" and "languages I'm able to express myself in". Some people are able to understand a language, but are not able to express themselves in it. We should also have the possibility to follow all the languages.

There is also the cases of multilingual toots. We should be able to choose more than one language when tagging manually a specific toot.

[jeanbaptisteb]

If this feature is implemented (which would be great), it may be relevant to use the ISO 639-3 code list (https://en.wikipedia.org/wiki/ISO_639-3).

This standard includes the possibility to tag a language as "undetermined", or to tag the content as not including text. Think for example about people just tooting a photograph (or a more complicated case: someone tooting just a link to a video with subtitles in multiple languages).

ISO 639-3 also includes codes for some artificial, rare, or ancient languages. Contrary to ISO 639-1 (for example), this standard tries to be as comprehensive as possible; it would be more inclusive to use it.

Besides the question of the standard to choose, another thing is to do is to separate the notions of "languages I'm able to understand" and "languages I'm able to express myself in". Some people are able to understand a language, but are not able to express themselves in it. We should also have the possibility to follow all the languages.

There is also the cases of multilingual toots. We should be able to choose more than one language when tagging manually a specific toot.

[jeanbaptisteb]

Another thing is that we should beware of a too exclusive filter system.

For example, toots without language (e.g. just a photograph, to take this example again) shouldn't be excluded from users' timelines, even these users chose to see toots only in Japanese or Spanish (for example).
Otherwise, toots without a defined language won't be seen by other users, except those who chose to see all languages.

[jeanbaptisteb]

Another thing is that we should beware of a too exclusive filter system.

For example, toots without language (e.g. just a photograph, to take this example again) shouldn't be excluded from users' timelines, even these users chose to see toots only in Japanese or Spanish (for example).
Otherwise, toots without a defined language won't be seen by other users, except those who chose to see all languages.

[Toub]

What about:

• don't exclude anything by default
• when a message is tagged (automatically or manually), show the language flag: EN, FR, ...
• when the user right clicks on the language flag, show an option in the menu: "Filter messages in this language"

Because you may want to filter messages if you receive too many of them (e.g. in French or English), but you will probably be happy (or don't mind) to receive other languages if it is just one message a week.

[Toub]

What about:

• don't exclude anything by default
• when a message is tagged (automatically or manually), show the language flag: EN, FR, ...
• when the user right clicks on the language flag, show an option in the menu: "Filter messages in this language"

Because you may want to filter messages if you receive too many of them (e.g. in French or English), but you will probably be happy (or don't mind) to receive other languages if it is just one message a week.

[Cassolotl]

I was asked to pass on a message here:

IDEA:
a per account/per instance language(s) tag that can be used to filter visible toots.

I really, really would like the federated timeline to be useable again.

[Cassolotl]

I was asked to pass on a message here:

IDEA:
a per account/per instance language(s) tag that can be used to filter visible toots.

I really, really would like the federated timeline to be useable again.

[pnathan]

This is going to be a problem as Mastodon increasingly goes global.

While I respect other users of Mastodon in Japan, Brazil, India, China, France, etc - and I wish them the best! - I can only read English and German. I can only scroll past toots which I can't read. I would like a way to, with the most polite respect, not see the toots I can't read when I am generally scrolling by.

It seems that a box where a user can set "my default toot reading/writing language is", and otherwise the user can override in their toot (perhaps under an advanced options dialog) if they want to read or write a toot outside of the defaults.

[pnathan]

This is going to be a problem as Mastodon increasingly goes global.

While I respect other users of Mastodon in Japan, Brazil, India, China, France, etc - and I wish them the best! - I can only read English and German. I can only scroll past toots which I can't read. I would like a way to, with the most polite respect, not see the toots I can't read when I am generally scrolling by.

It seems that a box where a user can set "my default toot reading/writing language is", and otherwise the user can override in their toot (perhaps under an advanced options dialog) if they want to read or write a toot outside of the defaults.

[NthTensor]

I think the simplest solution is to ask new users to select the tags when creating accounts, and use @Toub's suggestion for existing users. Essentially leaving the language filter blank to grandfather people in.

[NthTensor]

I think the simplest solution is to ask new users to select the tags when creating accounts, and use @Toub's suggestion for existing users. Essentially leaving the language filter blank to grandfather people in.

[jeanbaptisteb]

Too many features and options could make the interface more complicated, when the purpose of filtering by language is originally to make Mastodon more user-friendly.

For this reason, I think we should remove the "only show specific languages" option:

This option is a good idea, but it makes the preferences' interface a bit more complex. And the @Toub 's solution (basically a blacklist system) already solves the original problem of seeing too much toots we don't understand.

A language blacklist system is more inclusive and already takes into account complicated cases (images, links to multilingual content). Plus, a white list system has some problems (i.e. excluding content I could actually understand). Moreover, I fear that the coexistence of a whitelist and a blacklist systems would make the interface more complicated for users who are not tech-savvy.

So my preference would go to the implementation of a blacklist system exclusively.

Any thoughts?

About the ISO-3 languages, Wikipedia provides tranlations of their names in 6 languages: https://en.wikipedia.org/wiki/ISO_639%3aa . We could use it to make the interface more user-friendly. For example, in my case (I'm French), "Japonais" would be displayed instead of "Japanese" here :

When the translation of the language name isn't available in the user's language, then we could display its name in English.

[jeanbaptisteb]

Too many features and options could make the interface more complicated, when the purpose of filtering by language is originally to make Mastodon more user-friendly.

For this reason, I think we should remove the "only show specific languages" option:

This option is a good idea, but it makes the preferences' interface a bit more complex. And the @Toub 's solution (basically a blacklist system) already solves the original problem of seeing too much toots we don't understand.

A language blacklist system is more inclusive and already takes into account complicated cases (images, links to multilingual content). Plus, a white list system has some problems (i.e. excluding content I could actually understand). Moreover, I fear that the coexistence of a whitelist and a blacklist systems would make the interface more complicated for users who are not tech-savvy.

So my preference would go to the implementation of a blacklist system exclusively.

Any thoughts?

About the ISO-3 languages, Wikipedia provides tranlations of their names in 6 languages: https://en.wikipedia.org/wiki/ISO_639%3aa . We could use it to make the interface more user-friendly. For example, in my case (I'm French), "Japonais" would be displayed instead of "Japanese" here :

When the translation of the language name isn't available in the user's language, then we could display its name in English.

[sramkrishna]

The other issue is scrolling. On mastodon.cloud, the federated and even the local timelines are scrolling by so fast that I can't even read it. There are usually 'peak times' for various countries and denizens of those countries will be more chatty than others. Filtering will help at least in languages for those who cannot participate in languages we don't know. Unless of course, there is some on-the-fly translator.

When we hit peak time, the web client becomes very slow and unresponsive and I'm forced to mute everyone. That seems counteractive to me.

[sramkrishna]

The other issue is scrolling. On mastodon.cloud, the federated and even the local timelines are scrolling by so fast that I can't even read it. There are usually 'peak times' for various countries and denizens of those countries will be more chatty than others. Filtering will help at least in languages for those who cannot participate in languages we don't know. Unless of course, there is some on-the-fly translator.

When we hit peak time, the web client becomes very slow and unresponsive and I'm forced to mute everyone. That seems counteractive to me.

[Lomplac]

@jeanbaptisteb
Only showing the "Muted languages" option and having a "Show more options" button below would be a great way to allow the users have a user-friendly and complete interface that focus on inclusivity
Also it seems obvious that it's either the blacklist or the whitelist, the "muted languages" options would be disabled when one use the "only show specific languages" option

Showing the full name of languages on every toot would be unpleasant to read imo, if you want to know the language name, you just have to click on it ("Mettre le japonais en sourdine")

[Lomplac]

@jeanbaptisteb
Only showing the "Muted languages" option and having a "Show more options" button below would be a great way to allow the users have a user-friendly and complete interface that focus on inclusivity
Also it seems obvious that it's either the blacklist or the whitelist, the "muted languages" options would be disabled when one use the "only show specific languages" option

Showing the full name of languages on every toot would be unpleasant to read imo, if you want to know the language name, you just have to click on it ("Mettre le japonais en sourdine")

[lanodan]

Actually I think the thing should be flexible enough to allow both whitelist-style (I want only english and no-lang) or backlist-style (I don’t want to see german).
Also tagging by the browser just won’t work, manual tagging might work if there is a choice-list(text-form might be nice too for like non-recognised languages/dialects) based on previous use and automatic tagging done client-side (so the user can override) might be a good idea (users are lazy?).

[lanodan]

Actually I think the thing should be flexible enough to allow both whitelist-style (I want only english and no-lang) or backlist-style (I don’t want to see german).
Also tagging by the browser just won’t work, manual tagging might work if there is a choice-list(text-form might be nice too for like non-recognised languages/dialects) based on previous use and automatic tagging done client-side (so the user can override) might be a good idea (users are lazy?).

[ernix]

I think we only need a regex option for Local/Federated Timeline.
Filtering is a matter with viewers, we shouldn't make users to choose what language they are using.
Tagging language does not work for trolling and ads.

Regex filter is already implemented in Home.
Use [^ -~] to keep your timeline ascii.
Use .{200} to filter out long toots.

Simple, isn't it?

[ernix]

I think we only need a regex option for Local/Federated Timeline.
Filtering is a matter with viewers, we shouldn't make users to choose what language they are using.
Tagging language does not work for trolling and ads.

Regex filter is already implemented in Home.
Use [^ -~] to keep your timeline ascii.
Use .{200} to filter out long toots.

Simple, isn't it?

[jeanbaptisteb]

@Lomplac : Agreed.

About the way of showing the whitelist and blacklist in the "preferences" interface, what about a list of languages that you can check and uncheck, with options "select all" and "deselect all"?

The list of languages may be long, but we could show 10 main languages by default (for example), with an "expand" option for people who need to see more languages in the list.

I think it would be still possible to include the "temporary mute" feature with such an interface, the language would be temporarily unchecked for 1 hour for example, then will return automatically to its checked state.

"Showing the full name of languages on every toot would be unpleasant to read imo": agreed too.

@ernix : As far as language issues are concerned, I'm afraid your solution of filtering non-ascii characters doesn't work for people who are not English speakers (e.g. I'm French, I need to read non-ascii characters, but I want to filter toots in Japanese because I don't understand this language). Yet the filtering of timelines would be a great addition, maybe we should make a separate issue?

[jeanbaptisteb]

@Lomplac : Agreed.

About the way of showing the whitelist and blacklist in the "preferences" interface, what about a list of languages that you can check and uncheck, with options "select all" and "deselect all"?

The list of languages may be long, but we could show 10 main languages by default (for example), with an "expand" option for people who need to see more languages in the list.

I think it would be still possible to include the "temporary mute" feature with such an interface, the language would be temporarily unchecked for 1 hour for example, then will return automatically to its checked state.

"Showing the full name of languages on every toot would be unpleasant to read imo": agreed too.

@ernix : As far as language issues are concerned, I'm afraid your solution of filtering non-ascii characters doesn't work for people who are not English speakers (e.g. I'm French, I need to read non-ascii characters, but I want to filter toots in Japanese because I don't understand this language). Yet the filtering of timelines would be a great addition, maybe we should make a separate issue?

[ernix]

@jeanbaptisteb Yes, filtering by regex is another topic.

But when combining multiple filtering rules, the situation will be undesirable.
I wouldn't to be mad with filtering order and conditional filtering rules.

BTW, you can also allow diacritical marks in regex: [^ -~À-ÿ]

[ernix]

@jeanbaptisteb Yes, filtering by regex is another topic.

But when combining multiple filtering rules, the situation will be undesirable.
I wouldn't to be mad with filtering order and conditional filtering rules.

BTW, you can also allow diacritical marks in regex: [^ -~À-ÿ]

[Cassolotl]

I'm a normal user with non-technical knowledge, and I'm fine with regex being used as long as it's not the only way. I want to be able to see only French and English posts, but I know nothing about regex. Regex only would not be accessible for me.

[Cassolotl]

I'm a normal user with non-technical knowledge, and I'm fine with regex being used as long as it's not the only way. I want to be able to see only French and English posts, but I know nothing about regex. Regex only would not be accessible for me.

[jeanbaptisteb]

@ernix : @Cassolotl is right, it's not accessible for non-tech users. And anyway there are a lot of cases where you cannot filter languages with only regular expressions.

For example, if I toot the phrase "Il fait beau ce matin", an English speaker who uses regexp to filter non-ascii character will see my toot, even if it's not in English.

[jeanbaptisteb]

@ernix : @Cassolotl is right, it's not accessible for non-tech users. And anyway there are a lot of cases where you cannot filter languages with only regular expressions.

For example, if I toot the phrase "Il fait beau ce matin", an English speaker who uses regexp to filter non-ascii character will see my toot, even if it's not in English.

[ernix]

@Cassolotl @jeanbaptisteb

Have you ever set Content-Language or Accept-Language headers on your emails?
https://tools.ietf.org/html/rfc3282

I've never ever seen them before. Language tagging is totally useless.
Plus, everyone can setup dedicated instances for specific languages.

[ernix]

@Cassolotl @jeanbaptisteb

Have you ever set Content-Language or Accept-Language headers on your emails?
https://tools.ietf.org/html/rfc3282

I've never ever seen them before. Language tagging is totally useless.
Plus, everyone can setup dedicated instances for specific languages.

[Cassolotl]

As things stand, I'm sad that I can't change the language per-toot, because sometimes I post in French (very rarely, but I do!) and I know other people who post even half-English half-German or something. So I'm not sure how this is going to go down, but I'm a little worried.

@ernix I'm not really sure what you're talking about, so I can't take your point! Could you clarify?

[Cassolotl]

As things stand, I'm sad that I can't change the language per-toot, because sometimes I post in French (very rarely, but I do!) and I know other people who post even half-English half-German or something. So I'm not sure how this is going to go down, but I'm a little worried.

@ernix I'm not really sure what you're talking about, so I can't take your point! Could you clarify?

[jeanbaptisteb]

@ernix

We can make language tagging user-friendly: for example defining a language by default, then being able to change language of each toot manually for users who want to do so. Or even tagging languages automatically (I think it's already the case BTW).

I think it would be a terrible idea to force users to tag their toots, if it's what you're afraid of. I'm talking about making manual tagging an option, not about making it mandatory.

In any case, language tagging isn't useless for the large number of users who complained about being unable to filter by language.

[jeanbaptisteb]

@ernix

We can make language tagging user-friendly: for example defining a language by default, then being able to change language of each toot manually for users who want to do so. Or even tagging languages automatically (I think it's already the case BTW).

I think it would be a terrible idea to force users to tag their toots, if it's what you're afraid of. I'm talking about making manual tagging an option, not about making it mandatory.

In any case, language tagging isn't useless for the large number of users who complained about being unable to filter by language.

[ernix]

@Cassolotl
If you want to toot with language tags, you can just put language hash-tags like #ja #en #de #fr.
If you want me to toot with language tags, I'm sorry but I wouldn't.

@jeanbaptisteb
That's not friendly at all. I would rather prefer to add all language tags on every single toot so that I can avoid selecting tags each time. I don't want my toots to be tagged automatically.

As you said, there is no deterministic method to define what language the sentence is.
e.g. I'm a Japanese/Chinese speaker, "手紙" is a word that has two different meanings for each languages.

I need to keep my expression as is. You can judge it by yourself, but nobody can label it on specific language.

I hope this makes sense to you.

[ernix]

@Cassolotl
If you want to toot with language tags, you can just put language hash-tags like #ja #en #de #fr.
If you want me to toot with language tags, I'm sorry but I wouldn't.

@jeanbaptisteb
That's not friendly at all. I would rather prefer to add all language tags on every single toot so that I can avoid selecting tags each time. I don't want my toots to be tagged automatically.

As you said, there is no deterministic method to define what language the sentence is.
e.g. I'm a Japanese/Chinese speaker, "手紙" is a word that has two different meanings for each languages.

I need to keep my expression as is. You can judge it by yourself, but nobody can label it on specific language.

I hope this makes sense to you.

[jeanbaptisteb]

@ernix "That's not friendly at all. I would rather prefer to add all language tags on every single toot so that I can avoid selecting tags each time. ":

You could tag your toots as "language undetermined" by default, or not choose any language by default. But it doesn't mean that we have to forbid other users from tagging their toots if they want to.
I don't want you to tag your toots if you don't want to. But I think it would be great to give other users the ability to tag their toots if they want to, in the easiest possible way.

And I'm afraid that manually adding tags like #fr #de etc. isn't the easiest possible way. For example, non-technical users generally don't know what language codes would be the most appropriate to use (for example, who knows that the language code for Irish is "ga"?).

Thinking about making tagging easier is not incompatible with not making you obligated to manually add a tag if you don't want to. Again, I'm talking about giving an option to tag toots, not about making it mandatory.

[jeanbaptisteb]

@ernix "That's not friendly at all. I would rather prefer to add all language tags on every single toot so that I can avoid selecting tags each time. ":

You could tag your toots as "language undetermined" by default, or not choose any language by default. But it doesn't mean that we have to forbid other users from tagging their toots if they want to.
I don't want you to tag your toots if you don't want to. But I think it would be great to give other users the ability to tag their toots if they want to, in the easiest possible way.

And I'm afraid that manually adding tags like #fr #de etc. isn't the easiest possible way. For example, non-technical users generally don't know what language codes would be the most appropriate to use (for example, who knows that the language code for Irish is "ga"?).

Thinking about making tagging easier is not incompatible with not making you obligated to manually add a tag if you don't want to. Again, I'm talking about giving an option to tag toots, not about making it mandatory.

[Cassolotl]

@jeanbaptisteb @ernix Yes!

My ideal is an option in settings to set a default language, and you can leave it Unspecified.

And then in the "new post" interface, a language dropdown box, which is set to your default choice (for me English, but for ernix perhaps it would be Unspecified) - and if I write a post in French, I can use the dropdown to change it to French.

And then, in the various feeds and timelines, I would like to be able to filter to show only posts set to English, French, and Unspecified. But the default would be "view all posts in any language."

So everything is optional and Unspecified by default, and people can use hashtags like #fr if they want to. But there are user-friendly tools available that reduce work for people who want to be more specific.

[Cassolotl]

@jeanbaptisteb @ernix Yes!

My ideal is an option in settings to set a default language, and you can leave it Unspecified.

And then in the "new post" interface, a language dropdown box, which is set to your default choice (for me English, but for ernix perhaps it would be Unspecified) - and if I write a post in French, I can use the dropdown to change it to French.

And then, in the various feeds and timelines, I would like to be able to filter to show only posts set to English, French, and Unspecified. But the default would be "view all posts in any language."

So everything is optional and Unspecified by default, and people can use hashtags like #fr if they want to. But there are user-friendly tools available that reduce work for people who want to be more specific.

[ernix]

@jeanbaptisteb

You could tag your toots as "language undetermined" by default

Yes, I can. But why I have to do so?
I don't think it's effective to filter timelines when anybody can pick multiple languages.
There is no incentives for me to make my toots to be censorable by others.

Who knows that the language code for Irish is "ga"?

I do believe people who talk Irish language know the name, GAeilg.
Otherwise they can't select their own language from the Language list in Wikipedia.
https://en.wikipedia.org/wiki/Irish_language

@Cassolotl

And then in the "new post" interface, a language dropdown box, which is set to your default choice (for me English, but for ernix perhaps it would be Unspecified) - and if I write a post in French, I can use the dropdown to change it to French.

Yeah, It is completely OK that some client applications have features to append language hash-tags from drop-down UI. I don't care if these clients hide language hash-tags on the toots.

But I oppose system-wide language meta tag as @Lomplac said.

I don't want to see someone toot 六四事件(Six-Four incident) with zh(Chinese-Simplified) tag in order to attack Chinese people knowingly.

I also wonder why nobody commented with language tag on this thread.
You are talking "How language meta tags are useful" without language tags.

[ernix]

@jeanbaptisteb

You could tag your toots as "language undetermined" by default

Yes, I can. But why I have to do so?
I don't think it's effective to filter timelines when anybody can pick multiple languages.
There is no incentives for me to make my toots to be censorable by others.

Who knows that the language code for Irish is "ga"?

I do believe people who talk Irish language know the name, GAeilg.
Otherwise they can't select their own language from the Language list in Wikipedia.
https://en.wikipedia.org/wiki/Irish_language

@Cassolotl

And then in the "new post" interface, a language dropdown box, which is set to your default choice (for me English, but for ernix perhaps it would be Unspecified) - and if I write a post in French, I can use the dropdown to change it to French.

Yeah, It is completely OK that some client applications have features to append language hash-tags from drop-down UI. I don't care if these clients hide language hash-tags on the toots.

But I oppose system-wide language meta tag as @Lomplac said.

I don't want to see someone toot 六四事件(Six-Four incident) with zh(Chinese-Simplified) tag in order to attack Chinese people knowingly.

I also wonder why nobody commented with language tag on this thread.
You are talking "How language meta tags are useful" without language tags.

[jeanbaptisteb]

@ernix
"Yes, I can. But why I have to do so?". You wouldn't have to make any action. You could choose to not tag your toots. Then, the system would consider that no language has been defined for your toots.

You're talking about using hashtags like #fr, #ga, etc.

If I understand well your position, you're saying that users should be able to manually tag their toots with hashtags in order to enable language filtering, but at the same time that we shouldn't make it easy for them by implementing natively a way of doing it in the interface (with a button, for example).

I don't understand what is the argument in favor of this position, for me it doesn't make sense to implement filtering and then making it hard to use.

If you're against the idea of filtering, I think you should say it and to give your arguments for this. But for me, this discussion is about making a proposal to developers for implementing language filtering, taking into account use cases, yours included.

Not everyone will be happy in the end, but we have to acknowledge that the absence of language filtering is a problem for a large number of people. Personally, I won't discuss the relevance of language filtering with you, but I'm open to discuss how it could be implemented and how we could take into account your needs.

In the end, it's up to the developers of Mastodon to choose how and if language filtering will be implemented,.

[jeanbaptisteb]

@ernix
"Yes, I can. But why I have to do so?". You wouldn't have to make any action. You could choose to not tag your toots. Then, the system would consider that no language has been defined for your toots.

You're talking about using hashtags like #fr, #ga, etc.

If I understand well your position, you're saying that users should be able to manually tag their toots with hashtags in order to enable language filtering, but at the same time that we shouldn't make it easy for them by implementing natively a way of doing it in the interface (with a button, for example).

I don't understand what is the argument in favor of this position, for me it doesn't make sense to implement filtering and then making it hard to use.

If you're against the idea of filtering, I think you should say it and to give your arguments for this. But for me, this discussion is about making a proposal to developers for implementing language filtering, taking into account use cases, yours included.

Not everyone will be happy in the end, but we have to acknowledge that the absence of language filtering is a problem for a large number of people. Personally, I won't discuss the relevance of language filtering with you, but I'm open to discuss how it could be implemented and how we could take into account your needs.

In the end, it's up to the developers of Mastodon to choose how and if language filtering will be implemented,.

[ernix]

we shouldn't make it easy for them by implementing natively a way of doing it in the interface (with a button, for example).

Ah, this makes sense to me now. So you said user-friendly is about user accessibility on your browser, right?
You prefer to click/tap a tagging button rather than to type 3-4 keys.

I was thinking why dividing people into fixed language sets could be user-friendly. That's my mistake.

[ernix]

we shouldn't make it easy for them by implementing natively a way of doing it in the interface (with a button, for example).

Ah, this makes sense to me now. So you said user-friendly is about user accessibility on your browser, right?
You prefer to click/tap a tagging button rather than to type 3-4 keys.

I was thinking why dividing people into fixed language sets could be user-friendly. That's my mistake.

[lanodan]

[ @ernix ]

Have you ever set Content-Language or Accept-Language headers on your emails?
https://tools.ietf.org/html/rfc3282

Yes, I also receive many emails containing it. (probably because people are often using/setting languages pack/spelling)

Also for the issue of Japanese/Chinese : that was my point earlier, autocompletion of the language tag/header/… (people are lazy), but the user can set something else if it’s wrong/undetected.

[ @jeanbaptisteb ]

If you're against the idea of filtering, I think you should say it and to give your arguments for this.

I don’t really like idea of language filtering right now but few years back I really like that feature on friendica because I understood only english and french.
Also note that friendica UI’s put a spoiler/CW-like for posts in "foreign" languages.

[ @jeanbaptisteb ]

When the translation of the language name isn't available in the user's language, then we could display its name in English.

Actually, I prefer wikipedia’s way, original name in original writing system first, as when you know a language you often know how it’s written, but not how it’s written in other languages, and maybe user’s language and fallback to english.
English should not be so much "default" everywhere.
As said showing full-name would be horrible, ISO-codes are made for that, but full name would be useful when you toot. (As you can make a mistake between ISO-codes)

[lanodan]

[ @ernix ]

Have you ever set Content-Language or Accept-Language headers on your emails?
https://tools.ietf.org/html/rfc3282

Yes, I also receive many emails containing it. (probably because people are often using/setting languages pack/spelling)

Also for the issue of Japanese/Chinese : that was my point earlier, autocompletion of the language tag/header/… (people are lazy), but the user can set something else if it’s wrong/undetected.

[ @jeanbaptisteb ]

If you're against the idea of filtering, I think you should say it and to give your arguments for this.

I don’t really like idea of language filtering right now but few years back I really like that feature on friendica because I understood only english and french.
Also note that friendica UI’s put a spoiler/CW-like for posts in "foreign" languages.

[ @jeanbaptisteb ]

When the translation of the language name isn't available in the user's language, then we could display its name in English.

Actually, I prefer wikipedia’s way, original name in original writing system first, as when you know a language you often know how it’s written, but not how it’s written in other languages, and maybe user’s language and fallback to english.
English should not be so much "default" everywhere.
As said showing full-name would be horrible, ISO-codes are made for that, but full name would be useful when you toot. (As you can make a mistake between ISO-codes)

[yookoala]

Is there any room for mixed language contents? There are post on my timeline with Japanese, Chinese and English translation of the same content. Can a post be of Japanese, Chinese and English at the same time?

[yookoala]

Is there any room for mixed language contents? There are post on my timeline with Japanese, Chinese and English translation of the same content. Can a post be of Japanese, Chinese and English at the same time?

[ghost]

Please allow tagging toots with multiple language tags. I'm on an instance where several languages are used, and it's pretty common for people to include a translation into one or more other languages. Automatic tagging should also take this into account and add tags for all detected languages.

[ghost]

Please allow tagging toots with multiple language tags. I'm on an instance where several languages are used, and it's pretty common for people to include a translation into one or more other languages. Automatic tagging should also take this into account and add tags for all detected languages.

[ghost]

And please base the language detection on the full text of a toot, not just the part below a CW.

[ghost]

And please base the language detection on the full text of a toot, not just the part below a CW.

[jadykes]

This would be a terrific help; just about every toot in both timelines is now in japanese. I can't find toots I can read, which makes Mastodon completely unusable at times.

[jadykes]

This would be a terrific help; just about every toot in both timelines is now in japanese. I can't find toots I can read, which makes Mastodon completely unusable at times.

[Gargron]

This is now more or less in master.

[Gargron]

This is now more or less in master.

[jadykes]

Sorry, I'm new in GitHub - that means it's in the master...to-do list? And thanks for the swift response, wow! John

…

On May 20, 2017 6:16 PM, "Eugen Rochko" ***@***.***> wrote: This is now more or less in master. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#691 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AaO3US3_ff8FlQl38HCUhCL3LuWshuqQks5r73RfgaJpZM4MquZf> .

[jadykes]

Sorry, I'm new in GitHub - that means it's in the master...to-do list? And thanks for the swift response, wow! John

…

On May 20, 2017 6:16 PM, "Eugen Rochko" ***@***.***> wrote: This is now more or less in master. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#691 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AaO3US3_ff8FlQl38HCUhCL3LuWshuqQks5r73RfgaJpZM4MquZf> .

[jadykes]

Ahh, okay. Never mind!

…

On May 20, 2017 6:16 PM, "Eugen Rochko" ***@***.***> wrote: This is now more or less in master. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#691 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AaO3US3_ff8FlQl38HCUhCL3LuWshuqQks5r73RfgaJpZM4MquZf> .

[jadykes]

Ahh, okay. Never mind!

…

On May 20, 2017 6:16 PM, "Eugen Rochko" ***@***.***> wrote: This is now more or less in master. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#691 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AaO3US3_ff8FlQl38HCUhCL3LuWshuqQks5r73RfgaJpZM4MquZf> .

[nightpool]

@jadykes that means that it's been implemented

[nightpool]

@jadykes that means that it's been implemented