[PhilipRoman]

Some practical concerns I have with UTF-8 are similar (or even the same, depending on font) characters which can be used in malicious ways (think package names, URLs, etc), not to even mention RTL text and other control characters. Every time I add logging code, I make sure that any "interesting" characters are unambiguously escaped or otherwise signaled out-of-band. Having English as an international writing standard is perfectly fine and I say that as a non-native speaker with a non-ascii name.

[abdullahkhalids]

A good chunk of the world does not speak english or latin character based languages. They should be able to interact with computers completely in their own languages and alphabet sets, even if those are written right-to-left or top-to-bottom.

Of course, someone has to do the work to make this possible. And no one is obliged to do it. But to suggest that, such work should not be done at all, does not sit right.

[notpushkin]

This isn't quite black and white.

Right now, I can set up and use Linux in my language, have my display name in my script, but my username and password are ASCII-only and are available on the standard English keyboard anywhere. If I run into trouble, I can SSH in from any device in the world without any issue. I can just borrow a laptop from anyone, switch to English if needed, and jump right in.

Having a common denominator set of characters for such things is just really, really useful. I’d rather focus on all the other things that need to be localised.

[folmar]

Without any issue is a stretch, using a French keyboard is bad enough experience for passwords, not everyone uses standard English keyboards.

[wongarsu]

The French keyboard is the most notable example of anyone using something other than query or quertz. Even Japan and China use an extended querty. But even with the French keyboard the only issue is that everything is in the wrong place, not that the standard 26 "English" letters don't exist or are hard to reach.

Meanwhile using ä, é or ş in a username or password will make your life much harder once you are in a foreign country. Never mind any letter that isn't derived from the Latin alphabet.

[oarsinsync]

> something other than query or quertz. Even Japan and China use an extended querty

qwerty

[hnthrowaway6543]

> A good chunk of the world does not speak english or latin character based languages.

nearly everyone in a first world country knows the English alphabet though. a vast majority of the developing world as well. just look at street view on Google maps in any country, there's going to be a ton of street signs using English characters, even in non-touristy areas.

> They should be able to interact with computers completely in their own languages and alphabet sets, even if those are written right-to-left or top-to-bottom.

if you're a typical android/ios end user you're interacting with a computer in your native language anyway. this discussion only applies to low level power users.

in that case: why? these aren't user-facing features. this is like saying that people should be able to use symbols native to their language rather than greek letters when writing math papers.

it might not be "fair" that English is overrepresented in computing but it also hasn't demonstrably been a barrier to entry. Japan, Korea and China have dominated, particularly in hardware.

if you think it should be fixed why stop at usernames? why represent uids with 1234 instead of 一二三四?

[citrin_ru]

> nearly everyone in a first world country knows the English alphabet though

And not only 1st world. Actually the bigger country the more everything is localized - from dubbed films to food packaging labels. In a small country one would see more English/Spanish/French e. t. c. because they don't have resources to localize everything.

[abdullahkhalids]

> if you're a typical android/ios end user you're interacting with a computer in your native language anyway. this discussion only applies to low level power users.

I don't think you realize how poor this experience is. Partly the reason being that the underlying system is so english focused, that app developers have to do so much work to get things working.

> if you think it should be fixed why stop at usernames? why represent uids with 1234 instead of 一二三四?

I mean, if the computers had first been built in south east asia, they would have been.

[hnthrowaway6543]

it's certainly hard to localize everything but billions of people use ios/android in India, China, SEA, MENA, etc... i think it's fair to say that at the end user level, computers are in fact usable by non-English speakers.

individual apps may not be as usable, but that's on the developers. good counter-example, a lot of japanese games, even made within the past 5 years, require setting the Windows system locale to Japanese to function properly. and as someone who played a fair number of japanese doujin games in the 00s/10s, it used to be every game with this problem.

> I mean, if the computers had first been built in south east asia, they would have been.

debatable as CJK heavily use Arabic numerals everywhere, but even if they did, so what? you'd learn those symbols and get used to it. the same way that if you're a unix sysadmin you get used to only being able to use a small subset of ASCII characters for usernames.

[abdullahkhalids]

> it's certainly hard to localize everything but billions of people use ios/android in India, China, SEA, MENA, etc... i think it's fair to say that at the end user level, computers are in fact usable by non-English speakers.

Its important to contextualize these discussions in socioeconomics. Computers are not just fun play things. They are serious tools used for economic activities. Their usage, through their design, has significant impact on the social systems of society. Non-latin-language speakers are able to use poorly localized computers, but they are only able to use them less well than the latin-language speakers. At least in South Asia, there is a huge economic divide between those who can speak English and those who can't, where causality runs both ways, and in more recent times exacerbated by the inability of some to use technology. And that economic divide then causes huge sociopolitical problems in societies.

If computers are means for economic progress, we shouldn't put the condition that one has to somehow learn English to use them well. But isn't localization sufficient? No it isn't. Ignore even that localization requires some members of your language to be dual speakers. The current era of economic progress is characterized by software development. But if the only way you can develop software is to learn a foreign language, then surely we are denying economic progress to some communities.

P.S. I will repeat. Nobody has to do any work to help other communities. But to assert that such work should not happen is plain wrong.

[hnthrowaway6543]

you're confusing "speaking English" with "knowing the English alphabet." these things are orthogonal. 95%+ of people in those countries know the english alphabet. i just threw down google maps street view at a random spot in Phnom Penh and instantly found english letters visible from the street, on advertisements[0]. then i threw it down in a much smaller Thai city that i had never heard of, Nakhon Sawan, and instantly found English on the street.[1] i've been in China, Japan and Korea enough to know english characters are all over the place. the English alphabet is omnipresent everywhere, i think you fail to realize this. nobody who is using a computer in these places is getting confused by the english alphabet.

> But to assert that such work should not happen is plain wrong.

i assert it should not happen because it's not solving an actual problem, the same way that changing "x" and "y" to "ㅋ" and "ㅌ" in algebra doesn't solve a problem, and trying to "solve" it will yield to a monstrous amount of incompatibilities and confusion. here's a really good comparison: ipv6. IPv6 is solving a problem, maybe in a way people disagree with, but definitely a real problem... and yet we still can't make ipv6 fucking work after God knows how many years, and trying to get IPv6 networking at any sort of scale is a massive fucking headache. now we want to go through the same headaches to support... umlauts in usernames? yeah, no thanks.

there's enough real work left to be done in the world that we shouldn't waste time with stupid makework like this.

or maybe in 30 years i'll be able to call up IT support and say "hey i forgot my password, can you reset it? my username is 神王 سعود. ... need me to spell that for you?"

edit: somewhat ironically, HN swallowed a few of the unicode characters in my theoretical future username...

[0] https://i.imgur.com/0WkG0ze.png

[1] https://i.imgur.com/VhDR5Xh.png

[citrin_ru]

I have an impression that people confuse learning English (which is hard unless you native language is a Germanic/Romance one) with learning to recognize and type Latin characters which is easy and people around the world already use Latin alphabet without knowing any English. You may escape Latin alphabet if you have spend a whole life in a remote village but for people living in cities around the world it should be familiar and not a barrier at all. It's hard to escape Latin characters in the modern world and this ship has already sailed like it or not (I mostly do).

[Muromec]

Oh no please, I don’t want to have my linux username in Cyrillic. Thanks but no, thanks!

I know enough linux to see 10 ways in which it will make things worse at some point.

[smitelli]

> similar (or even the same, depending on font) characters which can be used in malicious ways

These are called "confusables" and boy does that well run deep: https://www.unicode.org/Public/security/16.0.0/confusables.t...