The apple and snowman are in Unicode because they were in legacy character sets, whereas TAFKAP's squiggle is a trade mark and does not appear in any character set I know of. If you think that the TAFKAP sign is required to representing some existing texts then you could make a fairly complicated Unicode character-inclusion proposal and argue for its inclusion in a future revision alongside soccer ball and baseball squared key.
The snowman is a bit odd, but the second character you presented, F8FF, which I'm guessing to be the Apple logo on your machine; is actually the last glyph in the private use area. So Unicode doesn't actually specify it.
Where can you register a unicode domain? I thought they were avoiding that for security reasons. To many glyphs look like others if you're not careful.
The domain name is xn--n3h.net, see http://tools.ietf.org/html/rfc3492 for details on how to encode unicode into your domain name. And yes this has been used for several fishing attacks.
The TLD registries are split into basically three categories
1. Those that didn't care for IDNs, and thus don't permit you to register them in their domain (of course they can't stop you using IDNs in sub-domains or host names, but those aren't felt to be a serious security problem, since you could already have hsbc.mydomain.example without IDNs and anyone falling for that probably lost all their money in phishing scams years ago)
2. Those that obeyed a directive to identify a set of characters suitable for their purpose and restrict IDNs to only those characters, while also determining and algorithm to prevent visual collisions, ie forbid different owners for domain names which would appear too similar when written. Various European ccTLDs are in this category, since they wanted to permit local characters in local business names, while avoiding security problems as much as possible.
3. Those that said "rules? we don't need no stinking rules, just give us your money". Most notably .com and other TLDs controlled by Verisign. These are the only registries where you can have arbitrary combinations of Unicode symbols in domain names. They're obviously a total security nightmare, but Verisign sees that as another opportunity to sell more SSL certificates. Sort of like selling fire insurance and poorly made fireworks in the same shop.
On my firefox I have a little box containing the number 26 above the number 03. I suspect that somewhere I'm missing some unicode something on my machine.
Safari on OSX gives me a nonplussed, fez wearing snowman, but Firefox 3.03 gives me a happy, top hat wearing snowman. After spending hours getting their ligatures down, the font designers seem to get a little punchy in the miscellaneous symbols.
One of the irritating things about IDN is that it doesn't seem to have spread to many email clients yet. Last time I tried mail.app, it % encoded ☭.net for me. Then complained that sending had failed. And then % encoded the % encoding when I tried again. Hilarity ensued.
Also fear of entering the US with ☭ on business cards.
Huh. Closer examination reveals that RFC4690 removes symbols from the list of permitted characters in domains. Verisign policy seems to be that existing domains can continue to be renewed but new ones can't be registered (something that seems to confuse most IDN registration sites), but there's no guarantee that browsers will continue to implement support for them.
Doesn't it seem unfair that Unicrud contains ☃ and but not
?
Heh. I'm guessing The Symbol is copyrighted. I wonder what the implications of that would be.
Surely it couldn't be more encumbered than .
The apple and snowman are in Unicode because they were in legacy character sets, whereas TAFKAP's squiggle is a trade mark and does not appear in any character set I know of. If you think that the TAFKAP sign is required to representing some existing texts then you could make a fairly complicated Unicode character-inclusion proposal and argue for its inclusion in a future revision alongside soccer ball and baseball squared key.
Well, I can't type the correct name of his 1992 album into iTunes, so I think that counts.
Nevermind that the album isn't very good.
It also seems like an oversight that http://www..com doesn't resolve.
The snowman is a bit odd, but the second character you presented, F8FF, which I'm guessing to be the Apple logo on your machine; is actually the last glyph in the private use area. So Unicode doesn't actually specify it.
There is an ☥ and a ♀, so it's halfway there.
Ah, I'd been sitting here confused, as each occurrence of that glyph (on my Linux box) looks like three horizontal lines.
I get Klingon. I suspect I need to find the responsible font and destroy it.
The font replies, "TODAY IS A GOOD DAY TO DIE"
Well, Linux uses its vendor-specific codepage for Klingon. Which says it all, really.
2603?
Bloat. ISO-8859-1 or die!
I hope your head falls off.
EBCDIC FOREVER!
I rock out with my baudot out
FIVE BITS WERE ENOUGH FOR HIM THEY ARE ENOUGH FOR ME .
Where can you register a unicode domain? I thought they were avoiding that for security reasons. To many glyphs look like others if you're not careful.
The domain name is xn--n3h.net, see http://tools.ietf.org/html/rfc3492 for details on how to encode unicode into your domain name. And yes this has been used for several fishing attacks.
The TLD registries are split into basically three categories
1. Those that didn't care for IDNs, and thus don't permit you to register them in their domain (of course they can't stop you using IDNs in sub-domains or host names, but those aren't felt to be a serious security problem, since you could already have hsbc.mydomain.example without IDNs and anyone falling for that probably lost all their money in phishing scams years ago)
2. Those that obeyed a directive to identify a set of characters suitable for their purpose and restrict IDNs to only those characters, while also determining and algorithm to prevent visual collisions, ie forbid different owners for domain names which would appear too similar when written. Various European ccTLDs are in this category, since they wanted to permit local characters in local business names, while avoiding security problems as much as possible.
3. Those that said "rules? we don't need no stinking rules, just give us your money". Most notably .com and other TLDs controlled by Verisign. These are the only registries where you can have arbitrary combinations of Unicode symbols in domain names. They're obviously a total security nightmare, but Verisign sees that as another opportunity to sell more SSL certificates. Sort of like selling fire insurance and poorly made fireworks in the same shop.
☂
☁
☀
godamnit. Last week I was trying to register the upside-down versions of my domains but was being declined :(
There's a list of more Unicode domains!
The snowman crashes Firefox (3.03, winXP), IE shows some of the unicode symbols, and opera mobile none of them.
I have FF 3.0.3 on xp and it works fine; might be an extension problem.
On my firefox I have a little box containing the number 26 above the number 03. I suspect that somewhere I'm missing some unicode something on my machine.
Safari on OSX gives me a nonplussed, fez wearing snowman, but Firefox 3.03 gives me a happy, top hat wearing snowman. After spending hours getting their ligatures down, the font designers seem to get a little punchy in the miscellaneous symbols.
I think I prefer the fez guy. It's a little less season-specific.
Apparently I'm using shitty fonts... here (Firefox 3.03, XP) it looks like hairy balls.
One of the irritating things about IDN is that it doesn't seem to have spread to many email clients yet. Last time I tried mail.app, it % encoded ☭.net for me. Then complained that sending had failed. And then % encoded the % encoding when I tried again. Hilarity ensued.
Also fear of entering the US with ☭ on business cards.
Huh. Closer examination reveals that RFC4690 removes symbols from the list of permitted characters in domains. Verisign policy seems to be that existing domains can continue to be renewed but new ones can't be registered (something that seems to confuse most IDN registration sites), but there's no guarantee that browsers will continue to implement support for them.
Firefox 3.0.3, Ubuntu 8.04, amd64, loads http://xn--n3h.net/ without issue.
Konqueror 3.5.10 attempts to load "http:?.net/", which fails.
I prefer the second behaviour. At least I'm safe from Phishing.
The fact that it displays the punycode http://xn--n3h.net/ rather than the snowman is supposed to be enough to protect you from phishing.
Read something like this.