Ask HN: Why are all non-English submissions flagged? | Hacker News Reader

Ask HN: Why are all non-English submissions flagged?

15 pointsNotSwift30 comments

For most people in the world, English is not their native language. Some articles, that contain interesting information, are only available in a different language. Yet, these articles are dead on arrival.

View on HN

Dang has mentioned quite often that the submissions should be in English, most recently https://news.ycombinator.com/item?id=27571809. If a submission in another language gains traction, it'll generally get changed to an English-language source instead, e.g. https://news.ycombinator.com/item?id=27922560.

How do you propose HN's anti-spam filters and Mod(s) check submissions? While I take your implied point about inclusiveness how many different languages do you expect the spam filters and Mods to be fluent in? One? Two? More? That opens you up to “What-About-ism”

In my opinion there is nothing wrong with a website starting up (and keeping topics and discussions) to their lingua franca. As a neutral-ish example – What if I went to a .ru forum and posted something Swahili? Chances are high that it would be flagged/deleted because Swahili is not a language they speak (the mods) and so they have no way of checking to see if it is a Viagra spam or a meaningful contribution.

Most browsers (i.e. all that are based on chromium) have integrated google translate allowing for automatic translation of websites, so no mod would have to learn the languages to filter out spam.

I'm not picking sides, but the same argument could be made for posting the link through Google Translate so that it appears English on HN. That way dang et al don't have to do it themselves in order to moderate.

Yeah, somewhere around the point HN became fifty percent submissions in languages I don't speak is the point where I would stop visiting at all. I'm not going to sort through all that looking for things I'll enjoy reading.

This isn't rocket science. HN is a website set up by English speakers and naturally therefore has a primarily English speaking readership. Any article not in English will only be of interest to a tiny minority of that audience.

So let me ask you, in turn, why you expect HN as a whole to be interested in non-English-language articles?

Back when I ran a fairly large forum we had to ban non-English content simply because we couldn't understand it, and so couldn't moderate it.

I'm guessing that's a big part of it - How do the HN moderators distinguish the credibility of a non-English news source? How can they determine if it's flame bait? How can they determine if it's legal or not?

If I see a submission in a foreign language, I'll paste it into Google Translate and flag if obviously off-topic.

Most of the non-English links I've checked are obviously off-topic, like "how to find a dentist in <city>".

I am not able to reproduce your observation, but I have only limited data. Could you give some more concrete examples of submissions that shouldn't have been flagged/killed?

Many of the times that I see a non-English link, it's spam, including completely non-HN-related links.

I went through the last ~25 hours of postings (with showdead enabled) to find foreign language links. What I found was that the dead links were to off-topic pages. Of the two non-dead pages, one in Spanish is mostly off-topic, and one in French is on-topic.

https://news.ycombinator.com/user?id=noakmilo has a couple in Spanish. It appears to be promoting their tech news site that isn't HN-related. Eg, "how to get a GMail account (as explained by NASA)." (That one wasn't dead.)

https://news.ycombinator.com/submitted?id=zmzm92 links in Arabic to a jobs site. Very off-topic.

https://news.ycombinator.com/user?id=revistacomebien posted two links in Spanish, including "Protein shakes that you should prepare for training". Off-topic.

anton96 posted the on-topic «Unix à jamais: nouvelle série en cours sur le Club iGen» at https://news.ycombinator.com/item?id=28078209 . gus_massa advised in the sole comment: 'post in other languages are usually ignored or flagged unless there is no equivalent information in English. At least translate the title to something like "Unix Forever: New Series Underway on Club iGen [in French]".'

I stopped at slightly more than 24 hours ago with a link by https://news.ycombinator.com/submitted?id=elizehurley which often links to a blogger entries with titles in Turkish like "Learn English Without Taking a Course".

Terrific!

I will note that after a while it must become self selecting though. If foreign language submissions never make it to the front page, eventually it will be rare that anyone bothers to submit the good stuff. Meanwhile the spammers will continue their attempts as usual.

Seems true they are flagged -

https://news.ycombinator.com/from?site=translate.google.com

This ones pretty interesting and flagged for instance, Western media didn't pick up the story so that's a loss to the community if they liked it -

Google project loon balloon fall in Brazil -

https://g1.globo.com/to/tocantins/noticia/2021/03/02/balao-c...

Quite disconcerting many commentators here seem to not be able to use Google Translate or know of it's existence.

You are right. A decent sized fraction of those Google Translate links don't seem out-of-place on HN had they been about the US. It makes me wonder if providing a GT link might be worse than linking directly to the non-English page (with an English description), then writing a comment linking to the GT page.

One dead link I would have likely have read is "Finland's “Broadband 4 everyone” is a total fuckup".

Thinking about it, I realized a downside of linking through Google Translate is that it's a bit more difficult to figure out if the domain name is a useful predictor of quality because it requires mousing over the URL instead of seeing the name. Perhaps that influences things?

> it's a bit more difficult to figure out if the domain

Shouldn't get it flagged by the rules though.

You could script the domain to be beside the original site and do a HN tag like 'Ask HN:' 'Translated:'

That's a lot of work that might be for nothing.

Different cultures and how they present the news brings in conflict to monocultures like HN (IT Nerds). People complaining about X has news ads with accidents on the side for instance.

It also allow loopholes to no News rules, since a tabloid murder might be big in country X but not in English.

It also would educate people in other cultures and things bigger than what the Western media feeds us. Bruce Schneider was wrong the other day because he used Western media, rather than jumping on cyber documents the Chinese government publishes.

And wonky translate is probably the world's largest written language after Engrish as the largest spoken language.

Hard to know where the stability would lie. I think it might add valuable information and is a small iteration. The information is hugely valuable, but might not be useable on HN.

Google Translate is fantastic for information. It is terrible for leisure reading. The vast majority of HN readers are here recreationally. For someone to want to read a Google translated page it has to both be important information but not so important there's a human-translated version of it. That doesn't leave a lot left.

> That doesn't leave a lot left.

It is high value information when only a few have it (I disagree that the majority of important things get human translated)

Using translate and dealing with people who's poor English is their second language is an important skill.

If HN is just recreation, so much so hard articles are flagged so not to interfere with that. That's ok.

If the information is that high value the submitter is free to translate it themselves. Readers don't have any more moral duty to machine translate and pick through a foreign language submission than they do to try and glean value from a submission written in terrible English by a teenybopper who doesn't use capitalization or punctuation and fills the page with emojis.

There are plenty of articles full of meme-speak that are intensely fascinating but I know if I want to share them with HN I'd better find or produce a version they'll find readable.

Then link to a translated version so that everybody can actually read it and not just a small subset of users.

Anecdotally, a very high percentage of non-English submissions are spam, scams, or self-promotion.

I guess primarily you should engage with local communities for that purpose.

At least in the past there were submissions via Google translate I believe to remember. If the article is particularly interesting that might be an option. I have no experience whether is still works.

Presumably because HN is an English speaking site?

Well there you go ladies and gentlemen, we are prepared to debate anything on HN, even whether we should allow non-english articles.

Bro, most of us can’t read other languages.

But for fun, given how no one really reads the articles, if the submitter translates it and summarized the gist, I think HN would oblige with a discussion. It wouldn’t be so bad so long as the these articles bring an entirely different experience from some place else in the world.

But I don’t need to read the same trite stuff in another language, there’s plenty in English.

I guess because most people come to hn are expecting to see stories mostly in English regardless of their nationalities.

why are most content-based community websites (or most websites really) single language?

They're created by someone and written in their native language, which in turn attracts a community that can communicate in that language.

why are most English websites English only? They are all created by the Brit?

Because most of the native English speakers are monolingual.

what point are you tring to make? Because mine is that it has nothing to do with "native". There are more English speakers, native and non-native, who are multilingual than speakers of any other language, yet they chose to make content in English, instead of multilingual to reach a wider audience.

በቀላሉ ብዙ ሰዎች ስለማይረዱ።

You can submit a link to Google translate.

Reply on news.ycombinator.com