I'm trying to understand the limitations of International Domain names... can someone explain what I can and can't do?
Can I register a neat domain name like e=mc2.com?
Are any of these symbols permitted: !@#$%^&*()_+?
I'm trying to understand the limitations of International Domain names... can someone explain what I can and can't do?
Can I register a neat domain name like e=mc2.com?
Are any of these symbols permitted: !@#$%^&*()_+?
Punycode only alters characters at codepoints greater than or equal to 128, therefore the punycode representation of
e=mc2
is stille=mc2
which can't be made a valid DNS name.Unicode is full of homographs, so the obvious thing to try would be to use U+FF1D (Fullwidth Equals Sign), unfortunately thanks to people registering paypal.com there are now controls on what characters are permitted, depending on the TLD/registrar (as another example, the new Russian Cyrillic TLD .рф disallows mixing Cyrillic and Latin character sets). The (currently proposed) RFC 5892 attempts to formalize rules for classifying which characters should be allowed in which cases.
For
.com
domains, see this page for documentation on which Unicode code points Verisign permit (and prohibit) in domain names.The general rule is that symbol characters are not permitted.
Looks like a-z, 0-9, and hyphens are allowed. As well as 1-63 characters long.
http://en.wikipedia.org/wiki/Hostname
https://www.rfc-editor.org/rfc/rfc952
https://www.rfc-editor.org/rfc/rfc1123
RFC 1035 is still the rule of the land when it comes to this. There are newer RFCs that explain how to encode unicode characters into so called LDH (lowercase, digits, hyphen) encoding. RFC 3492 describes punycode and 4690 summarizes the IDN landscape as a whole.
But to answer your question, only english alphabet letters, the digits from 0-9 and the hyphen character are allowed. Many DNS implementations also allow other characters such as the underscore "_".
Hopefully this helped, Rik