ICANN Under Pressure Over Non-Latin Characters 471
RidcullyTheBrown writes "A story from the Sydney Morning Herald is reporting that ICANN is under pressure to introduce non-Latin characters into DNS names sooner rather than later. The effort is being spearheaded by nations in the Middle East and Asia. Currently there are only 37 characters usable in DNS entries, out of an estimated 50,000 that would be usable if ICANN changed naming restrictions. Given that some bind implementations still barf on an underscore, is this really premature?" From the article: "Plans to fast-track the introduction of non-English characters in website domain names could 'break the whole internet', warns ICANN chief executive Paul Twomey ... Twomey refuses to rush the process, and is currently conducting 'laboratory testing' to ensure that nothing can go wrong. 'The internet is like a fifteen story building, and with international domain names what we're trying to do is change the bricks in the basement,' he said. 'If we change the bricks there's all these layers of code above the DNS ... we have to make sure that if we change the system, the rest is all going to work.'" Given that some societies have used non-Latin characters for thousands of years, is this a bit late in coming?
Re:not the whole internet! (Score:2, Interesting)
Seriously... How many mail servers are going to freak out because they can't handle unicode?
Can't trust your browser's address bar anymore. (Score:2, Interesting)
Unicode has many characters that look almost exactly like characters in Latin-1.
For example, if "www.microsoft.com" is shown in your browser's address bar, how would you know for sure that the "c" is not from the Cyrillic alphabet, or the "o" is not from the Greek alphabet?
You simply won't be able to trust your browser's address bar anymore. The possibilities for phishing attacks are endless.
Use a simple eight dot three kludge (Score:3, Interesting)
2. Doesnt require any change to the DNS system. (other than some name policy changes)
3. Allows links to be imbedded in normalweb-pages so that they can be cut and pasted by anyone with latin functionality. So a Japanese person could cut and paste the link to some arabic site that they dont have the font for.
4. While this is a kludge it has some major advantages over rebuilding the DNS system.
Storm
.cn (Score:3, Interesting)
domain in their local language. Leave
What's this going to do for security .. (Score:4, Interesting)
"A domain name is a unique address that allows people to access a website, for example, smh.com.au"
No,a domain name is a sequence of characters mapped to an IP address. It was designed so as you won't have to remember 66.35.250.150 instead of slashdot.org. This wasn't a problem while the original Internet consisted of just four computers. DNS was never designed to provide identity. There was also the case of a stock trader hacking a DNS server and redirecting traffic from a legitimate finantial site to his own where he had duplicated the real site only with bogus information.
"He said that this could create problems where, for example, a character in Urdu looks identical to one in Arabic"
It sure could. How about totally replacing DNS with a system of online identities.
Horrible indeed (Score:3, Interesting)
no wonder the middle east (arabic) countries are especially wanting this, because the majority of the inexperienced internet users there will be more likely to easily use these domain names, hence the sites using those domains will be greater incentive for controlling what they see, because these domains will be under their control nationally.
not only this, but we as it people will be very unwilling to change all our software to adapt with the new situation because of the horrible development/testing/implementation involved, and hence wont be accepting these domains as valid in our network traffic, which will create a second internet which is as described above, less free.
this should not be allowed.
Comment removed (Score:5, Interesting)
Bad for phishing (Score:3, Interesting)
Comment removed (Score:4, Interesting)
Re:Changing a system (Score:4, Interesting)
I don't have mediocre English spelling, and I would use the correct accented characters in English words like "naive" - except I don't know how to type those characters. Like many people, I know how to type the characters that are on the keyboard. Additionally, because there's no need for me to type characters outside the ones printed on the keys on my keyboard to make the internets come down my tubes, I have no incentive to learn how to type any differently than I already do.
It's not necessarily a matter of spelling ability.
The GNS System? (Score:5, Interesting)
Actually, DNS arguably is a giant search engine, which simply works on a 1:1 relationship and uses a distributed database (you input one piece of information, and it gives you some corresponding piece of information back). Replacing it with a 'fuzzier' search engine that would give you back a number of results, ranked by relevance, isn't that huge a leap.
Re:Changing a system (Score:1, Interesting)
http://www.krysstal.com/borrow_arabic.html [krysstal.com]
http://www.krysstal.com/borrow_farsi.html [krysstal.com]
http://www.krysstal.com/borrow_hebrew.html [krysstal.com]
HTH
Internet != Web, and other IDN technical issues (Score:3, Interesting)
The reason ICANN wants to do lots of testing (after having dragged their feet for years before getting started) is that IDNs fundamentally change how DNS works, and it's really important not to break too much when you do that (not that ICANN traditionally worried about that.) It's *not* simple, and you don't want to get it wrong.
DNS translates a set of strings of nominally-ascii characters into numbers, or translates numbers into a set of strings of characters, or translates some sets of strings into other sets of strings, depending on which query you run, and uses specific data formats to represent those strings and numbers. There are restrictions on what characters can be in the strings, some for reasons that we could easily declare to be obsolete (7-bit, uppercase-to-lowercase translation), some for reasons that are harder to change (printable characters only, please), and some which are really hard (dots are used as delimiters, and nulls terminate character strings in some popular computer languages. So you can't just plug in arbitrary Unicode two-byte characters instead of pairs of ASCII bytes and skip the case-munging, because some of the bytes will have values that can't be handled, though most of the 8-bit-character alphabets can be used transparently if you don't mind people using incorrect character sets on occasion. 8-bit character sets simply aren't enough - you can handle most Western languages in ISO-8859-1, and UTF-8 is closer but apparently not quite a cigar (too bad - it would have been my preference.)
The main IDN strategies replace this by adding one more translation layer - character-string-set IDN names are translated into ugly-but-recognizable Punycode strings, which get used with standard DNS character-string-set to number translations in the forward direction, and in the reverse direction, anything that arrived as a Punycode xn-uglystuff string usually gets fed to a Punycode-to-Unicode translator by a user interface.
Some things can be fixed by recompiling (or relinking, or re-DLLing) all of your programs with a DNS resolver library that guesses whether to convert strings or not - forward DNS knows to punycode non-ascii characters and not to re-punycode xn--uglystuff, though reverse DNS doesn't necessarily know whether to convert it to Unicode 16 or UTF-8 or just pass it on directly, and if you've typed in a domain name using something other than 7-bit lowercase+digits ASCII, it knows to punycode it, and obviously any domain registry supporting punycode ought to allow anybody who registers a name that doesn't need punycode to have both the straight and punycode names. But it's still ugly.