4 ms·

I still don't really understand why we had to use punycode for this instead of plain UTF-8? Unicode domains are nice, because it's always frustrating to have t

by seszett 10y ago

I still don't really understand why we had to use punycode for this instead of plain UTF-8? Unicode domains are nice, because it's always frustrating to have to misspell words just to make them fit within ASCII for no good reason (other than technical) but it really should have been handled in a better way, both on the technical side (punycode is ugly and scary for everyone) and the administrative side (it's actually not that difficult to handle phishing or to semi-automatically attribute domains with accents to the existing owners of the equivalent domains without accents, for example, at least for most European languages).

9point6 10y ago
I think part of the reasoning was down to there being a lot of Unicode characters that look identical to ASCII ones, and that it would make creating legitimate looking domains for phishing a lot easier.
- avian 10y ago
  My understanding is that punycode wasn't originally meant to be used by humans. It was just the workaround used to encode non-ASCII names in ASCII-only DNS records. Later browsers decided to show punycode in the URL bar instead of the decoded value in an effort to fight phishing until "they have a better solution". e.g. https://bugzilla.mozilla.org/show_bug.cgi?id=282270 https://bugzilla.mozilla.org/show_bug.cgi?id=282270
gkya 10y ago
«semi-automatically attribute domains with accents to the existing owners» But this is no good, as in many cases accented and not-accented words mean different things. In italian àncora means anchor and ancòra means still, why would àncora.it (say a big company producing naval anchors) get the rights to own also ancòra.it (which could be an interesting name for a blog, a political journal, etc.)?
- seszett 10y ago
  That's the semi part. I know there are some collisions, but for the most part there's no problem. It's really not that difficult to check with a dictionary, a national registry of companies, and contact the owners in case of ambiguity.

by seszett 10y ago

9point6 10y ago

I think part of the reasoning was down to there being a lot of Unicode characters that look identical to ASCII ones, and that it would make creating legitimate looking domains for phishing a lot easier.

avian 10y ago
My understanding is that punycode wasn't originally meant to be used by humans. It was just the workaround used to encode non-ASCII names in ASCII-only DNS records. Later browsers decided to show punycode in the URL bar instead of the decoded value in an effort to fight phishing until "they have a better solution". e.g. https://bugzilla.mozilla.org/show_bug.cgi?id=282270 https://bugzilla.mozilla.org/show_bug.cgi?id=282270

gkya 10y ago

«semi-automatically attribute domains with accents to the existing owners» But this is no good, as in many cases accented and not-accented words mean different things. In italian àncora means anchor and ancòra means still, why would àncora.it (say a big company producing naval anchors) get the rights to own also ancòra.it (which could be an interesting name for a blog, a political journal, etc.)?

seszett 10y ago
That's the semi part. I know there are some collisions, but for the most part there's no problem. It's really not that difficult to check with a dictionary, a national registry of companies, and contact the owners in case of ambiguity.