zlacker

[parent] [thread] 38 comments
1. c0balt+(OP)[view] [source] 2025-10-09 19:39:38
Interesting article, I've had some similar (though significantly less severe) experiences with having ä and ß in my names, it seems many U. S. companies are just unwilling/incapable of going beyond ASCII.

The government being this sloppy at getting accents right is surprising, I would expect them to value accuracy and a clean paper trail when handling names.

http://archive.today/5h4v2

replies(7): >>ninala+S3 >>comrad+n4 >>Y_Y+k6 >>Aloisi+gp >>6SixTy+ly >>Andriy+Cz >>Rendel+OB2
2. ninala+S3[view] [source] 2025-10-09 20:02:58
>>c0balt+(OP)
> The government being this sloppy at getting accents right is surprising, I would expect them to value accuracy

That tells me you're German, I didn't even need to see the ä and ß.

Even in the UK I encounter websites that won't accept my Norwegian address because it begins with Å. English speaking countries generally are pretty bad at this sort of thing.

replies(3): >>dcmint+A4 >>reorde+6b >>Quantu+Ki
3. comrad+n4[view] [source] 2025-10-09 20:05:37
>>c0balt+(OP)
The USA government can't even handle ü. I was filling out a simple form to replace my damaged passport. I live in Zürich but it couldn't handle the umlaut. I never know what to do in this situation - do I use 'ue' instead, which is most common in Europe, or do I just use 'u' which is wrong but usually works in America. I didn't even bother checking with 'ue' and just went with 'u'

Ü isn't even a special character or utf-8 - ü is part of ascii. How does this even fail? Is their database a 7-bit database?

replies(9): >>pixl97+Y4 >>nemoma+p5 >>techno+v5 >>immibi+p8 >>integr+F9 >>daemon+Ca >>cridde+tb >>umanwi+ti >>Mister+hl
◧◩
4. dcmint+A4[view] [source] [discussion] 2025-10-09 20:06:36
>>ninala+S3
Last time I tried to make an international transfer from my British bank account it couldn't accept destination names with any accented characters. An international transfer. Sheesh.
◧◩
5. pixl97+Y4[view] [source] [discussion] 2025-10-09 20:08:07
>>comrad+n4
Don't think of government ascii as the entire ascii code page. Instead think of it as what's on a typewriter. You get 0-9 and A-Z along with a few punctuation characters.
replies(1): >>krior+ku
◧◩
6. nemoma+p5[view] [source] [discussion] 2025-10-09 20:10:36
>>comrad+n4
if they can't type it on a keyboard when manually copying it from one place to another they'll drop it is what I've come to expect.
◧◩
7. techno+v5[view] [source] [discussion] 2025-10-09 20:10:48
>>comrad+n4
Heck, they can't even handle that my last name has more than one capital letter in it.
8. Y_Y+k6[view] [source] 2025-10-09 20:15:00
>>c0balt+(OP)
I know these characters now have lives of their own, but they emerged as shorthands for "ae" and "ss" respectively (where the first s was the long kind that looks like an 'f').

My understanding is that they are still phonetically entirely equivalent. How does it feel to have to substitute them into your name? (Or do you have a different recourse?)

replies(1): >>c0balt+Tp
◧◩
9. immibi+p8[view] [source] [discussion] 2025-10-09 20:24:46
>>comrad+n4
ü is not part of ascii. Are you thinking of latin-1?
replies(1): >>comrad+s9
◧◩◪
10. comrad+s9[view] [source] [discussion] 2025-10-09 20:31:23
>>immibi+p8
Oh that makes sense. ASCII is 7-bit. so they could be depending on old 7-bit databases.
replies(1): >>db48x+cd
◧◩
11. integr+F9[view] [source] [discussion] 2025-10-09 20:32:39
>>comrad+n4
ASCII is a 7-bit code. Ü is not a part of ASCII any more than say Ł is (it's in one of many 8-bit extensions of ASCII)
◧◩
12. daemon+Ca[view] [source] [discussion] 2025-10-09 20:37:49
>>comrad+n4
There's no ü in ASCII (the 1967 US-ASCII everyone thinks of, anyways, which is all you can expect from the government). It's in ISO 646:CH though, where it replaces '}', and in Latin 1.
◧◩
13. reorde+6b[view] [source] [discussion] 2025-10-09 20:40:12
>>ninala+S3
You'd think the UK would be alright with it given Irish names are and were reasonably common (at least in parts of the UK), which commonly have fadas in them (i.e. Béibhinn, Seán, Ciarán, etc)
replies(1): >>miki12+eC
◧◩
14. cridde+tb[view] [source] [discussion] 2025-10-09 20:42:00
>>comrad+n4
'u' isn't wrong in America. We spell it Zurich. And in China, it's 苏黎世 (or at least that's what Google tells me).
◧◩◪◨
15. db48x+cd[view] [source] [discussion] 2025-10-09 20:53:24
>>comrad+s9
No, they can only use what shows up on the keyboard. Internally the software is a vast mix of systems that in practice can probably handle unicode just fine by now. It's just that the people can't type any of those characters.
replies(1): >>Quantu+9k
◧◩
16. umanwi+ti[view] [source] [discussion] 2025-10-09 21:26:37
>>comrad+n4
> Ü isn't even a special character or utf-8 - ü is part of ascii

That is not true. Type “man ascii” on macOS or Linux to see everything that is part of ascii.

◧◩
17. Quantu+Ki[view] [source] [discussion] 2025-10-09 21:28:32
>>ninala+S3
Just because we have æ, ø, å here in Norway doesn’t mean that we’re much better at handling “weird” characters from other countries than the English speaking countries are.

My Spanish girlfriend has an ñ in her last name, and as does our son. To the people here in Norway, I just tell them to put a plain n when typing the last name. It’s easier to just go with that than to try and get people to understand how to type ñ on the keyboard (even though our computers can do it), and to avoid extra back and forth with people who have systems that don’t handle it.

Likewise, when I’m in Spain I don’t bother to say that my last name has ø in it. I don’t even bother to rewrite the o in my last name as oe. I just put it as o.

The only situation where I put it as oe is indirectly when an airline converts ø to oe on my airline ticket, or where the airline system doesn’t handle ø and I put it as oe for them when making the booking. To me my name looks worse with oe in it, and seems harder to pronounce for people if I write it as having oe in it than just putting it as o.

replies(1): >>ninala+CGf
◧◩◪◨⬒
18. Quantu+9k[view] [source] [discussion] 2025-10-09 21:38:45
>>db48x+cd
Option plus u on US English keyboard on macOS gives you the umlaut, and then hit u again and you have ü.

But I wouldn’t bother memorising that and every other possible way that the other person has to press the keys depending on their keyboard layout and operating system. I’d just tell people to put u instead.

replies(2): >>adastr+ED >>db48x+tr3
◧◩
19. Mister+hl[view] [source] [discussion] 2025-10-09 21:48:11
>>comrad+n4
> Is their database a 7-bit database?

Ascii is 7 bits. What people think of as 8-bit ASCII is actually code page 437, the alternate characters added to the PC BIOS in the original IBM PC. Like UTF-8 it uses the most significant bit in a 1 byte ASCII char to determine if it should use a character from ASCII if 0 or the extended 437 characters which includes ü if 1. https://en.wikipedia.org/wiki/Code_page_437

replies(1): >>umanwi+qD
20. Aloisi+gp[view] [source] 2025-10-09 22:24:45
>>c0balt+(OP)
A clean paper trail is only possible if everyone always enters your name in exactly how you spell it. That's not realistic though.

If you try spelling your name over the phone to an American government employee, the vast majority would have no idea what a eszett was or how to enter it. Even if you wrote ß on a form, most wouldn't be able to enter it. Nor would most know how to pronounce it.

Even for accented letters like ä which at least have a form someone might recognize, the sheer number of different accent marks used across languages and the difficulty in reading someone's handwriting and general unfamiliarity with foreign names is just asking for some clerk to enter in wrong.

And that's just names with Latin letters. It becomes infinitely worse once you start including all the other character in world languages.

Instead, US government databases usually have first and last names transliterated into uppercase non-accented letters and they match against the transliterated name. Middle names are often only for display purposes. If you're lucky, they'll be display versions of first and last as well where you might sometimes be able to stick an accented character.

This isn't really limited to the US either. If you look at any passport, you'll notice the machine-readable section does the exact same thing, so on German passports ß becomes SS and Ä becomes AE.

◧◩
21. c0balt+Tp[view] [source] [discussion] 2025-10-09 22:32:15
>>Y_Y+k6
> How does it feel to have to substitute them into your name?

It does not directly bother me but can lead to downstream inconveniences. Public services (in Germany) ime don't like mismatches in identifiers, especially inconsistent ones. If it is required then it might sometimes take more than one application (with a small explanation on why the mismatch is there).

As another example, if ä is substituted for ae in shipping addresses then automatic tracking for packages by DHL via my customer account breaks (as the address is not identical anymore).

◧◩◪
22. krior+ku[view] [source] [discussion] 2025-10-09 23:16:55
>>pixl97+Y4
Not very helpful - my typewriter has an ü :)
replies(2): >>pixl97+pv >>marcel+Yw2
◧◩◪◨
23. pixl97+pv[view] [source] [discussion] 2025-10-09 23:27:00
>>krior+ku
And I'm sure Chinese typewriters have a totally different charset then German ones.
24. 6SixTy+ly[view] [source] 2025-10-10 00:02:15
>>c0balt+(OP)
Don't completely rule out EBCDIC from the equation. It's generally an IBM-ism meant for punched cards with the same characters as ASCII. Despite the similarity, it means that it's a legacy system kept alive to now and switching to UTF-EBCDIC would likely incur disturbing 70+ year old layers of tech debt dust. Some of them don't even support lower case characters as it's using EBCDIC's predecessor, BCD.
replies(1): >>miki12+AB
25. Andriy+Cz[view] [source] 2025-10-10 00:17:25
>>c0balt+(OP)
It's a question of scope, though.

Imagine you are an American designing a system. What about non-Latin alphabets? Yeah, these should probably be converted, nobody's going to bother with those. What about Hungarians, should we care about their O / Ó / Ö / Ő and U / Ú / Ü / Ű? And Icelanders - should we allow their Ð / Þ?

I understand that seeing your name misspelled hurts, but pretending ASCII is enough for everyone is an understandable simplification.

replies(1): >>int_19+pP
◧◩
26. miki12+AB[view] [source] [discussion] 2025-10-10 00:40:35
>>6SixTy+ly
And this is even the case in Europe.

There was a case of some German bank treating ü as "ue", its typical ASCII transliteration. A customer complained under GDPR and won.

◧◩◪
27. miki12+eC[view] [source] [discussion] 2025-10-10 00:50:12
>>reorde+6b
I mean, English itself has diacritics in perfectly ordinary words that many Americans use every day (café, résumé, and yet somebody decided to put "useless" characters like @, ^ or # in ASCII instead.
replies(2): >>Tinker+WF >>int_19+kP
◧◩◪
28. umanwi+qD[view] [source] [discussion] 2025-10-10 01:04:55
>>Mister+hl
Do people think of this as 8-bit ASCII ? I've never heard of it referred to ASCII until now. In fact, I've never heard of it at all (by the time I was old enough to know what a character encoding was, Latin-1 and Windows-1252 were totally dominant IIRC).
replies(1): >>int_19+JO
◧◩◪◨⬒⬓
29. adastr+ED[view] [source] [discussion] 2025-10-10 01:09:26
>>Quantu+9k
This is probably a Dosbox system from the 90’s, with COBOL batch processing in the backend.
◧◩◪◨
30. Tinker+WF[view] [source] [discussion] 2025-10-10 01:44:53
>>miki12+eC
They are optional though, and who can be bothered to add that when typing.
◧◩◪◨
31. int_19+JO[view] [source] [discussion] 2025-10-10 04:07:04
>>umanwi+qD
In DOS days, codepage 437 was commonly referred to as "ASCII" colloquially in countries where some other encoding was in use for the national script.
replies(1): >>umanwi+LR
◧◩◪◨
32. int_19+kP[view] [source] [discussion] 2025-10-10 04:13:38
>>miki12+eC
^ comes from typewriters, where it was used to type letters like â by backing the caret and typing over the last letter. Similarly ' could be used for acute accent - hence why it didn't get a separate letter - and ` for grave accent.

By the way, the accents can often be used to force the right pronunciation of a foreign name on native speakers (at least in US, where Spanish names are so widespread). So e.g. use "á" if you want it to be pronounced [a] etc.

◧◩
33. int_19+pP[view] [source] [discussion] 2025-10-10 04:14:34
>>Andriy+Cz
ASCII isn't even enough for English.

And yes absolutely we should bring Ð / Þ back for English use and drop those ridiculous digraphs.

◧◩◪◨⬒
34. umanwi+LR[view] [source] [discussion] 2025-10-10 04:53:38
>>int_19+JO
Interesting to know, thanks!
replies(1): >>Mister+UL1
◧◩◪◨⬒⬓
35. Mister+UL1[view] [source] [discussion] 2025-10-10 13:43:38
>>umanwi+LR
Yup, what the GP said. It was a common mistake for people to assume ASCII included CP 437 characters back in the DOS days. Even I made that assumption when I was a kid.
◧◩◪◨
36. marcel+Yw2[view] [source] [discussion] 2025-10-10 17:59:25
>>krior+ku
But do American English typewriters have a ü?
37. Rendel+OB2[view] [source] 2025-10-10 18:27:29
>>c0balt+(OP)
I saw online that people moving to Japan can sometimes take months to get a bank account because the systems aren't there designed to handle middle names. If you want to maximize your child's ability to move around, I guess go with an ASCII first name-last name, which works almost everywhere. Except Iceland!
◧◩◪◨⬒⬓
38. db48x+tr3[view] [source] [discussion] 2025-10-11 00:40:35
>>Quantu+9k
You can’t expect a bureaucrat to know about that kind of thing! If it’s not printed on a key right in front of their nose then no government employee is going to be able to type it in.

But that’s not really the point. No matter how many keyboard shortcuts the clerk at the DMV memorizes there is always going to be some text that they just cannot reproduce accurately. Whether it’s an accented character from the exotic land of Spain or some real Zalgo, something is going to get lost. No individual human can correctly deal with all possible textual forms.

◧◩◪
39. ninala+CGf[view] [source] [discussion] 2025-10-15 13:44:06
>>Quantu+Ki
I'm not talking about ordinary members of the public not bothering to figure out how to type an unusual character on a Norwegian keyboard. I have no idea how I would type a an n with a tilde, I would have to pick it from the character map.

My point was that someone who can type it will often have it rejected by a website. I was using a hotel booking site and when I booked a room it asked me for my address so I typed Å... The web page rendered it correctly but when I hit the button to complete the transaction it told me that my address contained an illegal character (or some similar wording). And this site handles bookings for hotels that themselves have names with umlauts, tildes, cedillas, etc.

[go to top]