Initial non-ISO646 character refused in display name

Question

I tried to set my display name to:

نصر الدين تهريب الحمير

That is composed entirely of letters and spaces, Unicode-wise.

Yet the profile editor rejected it on the grounds that it wasn't composed of letter and numbers and/or didn't start with a letter or number.

So, either the message needs to say, 'starts with a number or a latin alphabet letter', or the checker needs to changed.

Note that this question has nothing to do with whether non-Latin names are permitted. It was easy to stick a zero on the front and use it.

The Bidi bugs that this is exposing are a bit of an extra bonus.

edit

Full disclosure. This is not my name. I'm not complaining about freedom of expression. If the site changes to reject نصر الدين تهريب الحمير, I have no cause for complaint. However, accepting <0 نصر الدين تهريب الحمير> and rejecting without the zero seems a bit odd, and the error message is, I claim, certainly inaccurate.

Google translate will tell you everything you need to know about this string. It goes with my new gravatar.

The Chinese string should be self-evident when Google-translated. For the Arabic, if you need help, post a question and I'll be happy to type in all the details in an answer.

NOTE

This idea of RTL display names seems to have been in the water: see this.

I don't think that this should be closed as a dup of that, since I'm noting the problem with the profile form validation, not the display problems that ensue.

Answer 1

we were not allowing anything other than [A-Z0-9] when validating the first character. Fixed.

Now we use ^[\w](?<!_)

Answer 2

But see what you've done…

Answer 3

Jeff has repeatedly said that he is not interested in globalizing/localizing any of the Trilogy sites.

As the question Is English required on Stack Overflow? (an official FAQ entry) says, using English as the primary (only) common language facilitates communication -- which is what these sites are all about.

How does putting your display name in Arabic facilitate communication with the population of an English-only site?

---

Edit: Okay, I misunderstood.

I think the reason the checker works like that is because some users have names that contain non-English characters, yet are still closely related enough to English characters (say, é) that they can be matched in features like comment replies.

The issue is that if some non-English characters are allowed, either the set of allowed characters must be explicitly included; or, the set of disallowed characters must be explicitly excluded -- neither of which is a very palatable choice.

I think this will probably go down as [status-bydesign] where display names like that will be dealt with on an individual basis, as it doesn't happen often.