| 2:21 pm on Oct 26, 2009 (gmt 0)|
So shall I now have to defend my names in non-Latin languages?
| 3:34 pm on Oct 26, 2009 (gmt 0)|
I suppose yes. If you own example.com, you'll soon be getting phishers from éxample.com, examplḛ.com, examƥle.com, examp³e.com, exaḿple.com, etc. In fact if you want to protect all the phishy versions of a typical 7-letter latin domain, you will need to register 3,229,832 new domains (approximately).
Renewal fees will be about $16M annually. Maybe your registrar will give you a bulk discount?
OUCH! all those special characters went kaboom. WebmasterWorld isn't UTF-8, huh
| 3:43 pm on Oct 26, 2009 (gmt 0)|
To quote general McAuliffe: "Nuts!"
| 4:02 pm on Oct 26, 2009 (gmt 0)|
I'm hoping this will be regulated so that domain names can either be all latin, all chinese, all hebrew, etc. Mixing it up will create enormous issues as outlined by HWW above.
| 4:10 pm on Oct 26, 2009 (gmt 0)|
|I'm hoping this will be regulated so that domain names can either be all latin, all chinese, all hebrew, etc. |
Right. What happens if they're all German, all French, all Norwegian, all Czech etc. ?
I'm very much in favour of the orthographic democratisation of web addresses, but I can't see how the potential tide of abuse can be controlled without a human-staffed domain-spam-watch taking reports in from webmasters who see phishy versions of their own domains. And that's just not practical.
[edited by: ronin at 4:13 pm (utc) on Oct. 26, 2009]
| 4:10 pm on Oct 26, 2009 (gmt 0)|
Surely the best way to implement this is on a local based setting? No Chinese characters, unless the domain is a ".cn" and no German special characters unless it is a ".de"
Otherwise it will end in all kinds of complications, such as the example.com examples by httpwebwitch.
I didn't get the impression that was going to happen from the article though?
| 4:34 pm on Oct 26, 2009 (gmt 0)|
The solution is simple - a map of similar characters is required so that similar domain names cannot be registered. Of course, this concept may be too advanced for ICANN to fathom.
| 4:50 pm on Oct 26, 2009 (gmt 0)|
There seems to be some confusion here, httpwebwitch and others you can currently buy .com / .net / .org and others using non-latin characters.
For example httpwebwitch's example of: éxample.com is owned by someone since 2006 and is a working site. There are many of these types of sites all over the Internet. One of the names that I like is: £.com (yes its a working site)
Phishing is a big problem, you can actually buy 100% visually matching domains such as any domain that uses an: a o e and other letters are easily phished with links (even from AdWords) with domains that look 100% perfect but enter them into a broswer and you are taking to other websites.
What ICANN is looking at doing is allowing the TLD part of the domain (.com/.net/.org) to be non-latin. This is like what you get places like China that allow 100% Chinese domain names.
So while people are currently trying to get TLD's like .sex you can if the above happens put in the request to have them setup a .séx TLD
I welcome the idea, but it will get very confusing for some users due to most countries now either becoming used to the .com version of the CCTLD version of a website.
As to date there hasn't been any widespread success with a TLD other then .com but I do see in Europe the .eu coming more and more into play, so it might have a future?
| 5:17 pm on Oct 26, 2009 (gmt 0)|
|Right. What happens if they're all German, all French, all Norwegian, all Czech etc. ? |
they continue using standard latin characters as they do now. Without accents. If accented latin characters are allowed, it will be a phishing nightmare. So ICANN should state that a domain must be written all-latin (like now, without accents) or all using other alphabets like chinese, japanese, thai, etc.
| 5:22 pm on Oct 26, 2009 (gmt 0)|
However it does resolve to poundsymbol . com, is that allowed Mods?
| 6:03 pm on Oct 26, 2009 (gmt 0)|
My take, is the news is the use of the word "entire" in the following quote:
|One of the key issues to be taken up by ICANN's board is whether to allow entire internet addresses to be in scripts that are not based on Latin letters. |
Second level (and sub) domain portions of an address can already use IDN characters (as others have pointed out).
What "entire" suggests is non-ASCII TLD's... and may bring up issues like; "What is a dot?"..
| 6:15 pm on Oct 26, 2009 (gmt 0)|
BTW - the "pound" site appears to be a well planned domain buy -- they needed to own both the symbol and the "poundsign" domains -- the translation from the IDN character domain is done through punycode (to ASCII), as:
£.com --> xn--9a.com --> poundsign.com
Use the Verisign IDN Conversion Tool to make sense of all this:
| 10:55 pm on Oct 26, 2009 (gmt 0)|
|calling it a "fantastically complicated technical feature". |
That doesn't engender confidence.
| 12:36 am on Oct 27, 2009 (gmt 0)|
This has the potential to break many older programs which do not support multi-byte character types.
| 1:45 am on Oct 27, 2009 (gmt 0)|
Well that's rubbish for a start. The .de ccTLD is roughly 13.2 million domains. The .cn ccTLD is around 13M and the .uk ccTLD is around 7.8M. The .eu "ccTLD" is a disaster when it the registrations are broken down over all 27 member states of the EU. ( [eurid.eu...] ) The reality is that the EU is very much a ccTLD area where the ccTLD domain count is greater than or close to that of .com domain registrations in those countries. In terms of market relevance, the .eu is competing with .org, .info and .biz in most of the EU markets. It is only in the Eastern European states that it has had anything that could be called a success.
|As to date there hasn't been any widespread success with a TLD other then .com but I do see in Europe the .eu coming more and more into play, so it might have a future? |
These new gTLDs and the non-Latin character TLDs are, in my opinion, an attempt to keep ICANN in what is increasingly becoming a more ccTLD orientated game.
| 7:05 am on Oct 27, 2009 (gmt 0)|
|The solution is simple - a map of similar characters is required so that similar domain names cannot be registered. |
Actually that solution is very complicated. There are umpteens of ways to transliterate a name from Russian or Arabic or Chinese etc. to Latin characters.
| 9:34 am on Oct 27, 2009 (gmt 0)|
If a letter looks like an "e" (e.g. it has an accent) treat it as an "e". Follow this though for all unicode characters and the problem is solved.
This is no more complicated conceptually than mnemonic phone numbers - each digit represents several letters but each number must be unique. The same must be true for domain names - all characters that appear similar must be treated as being equal for domain name registration.
There are several complications
1) Combinations : e.g. "rn" typically looks like "m" (this could be extended phonetically).
2) Human judgement : disagreements are bound to occur.
3) Case and font : probably upper case will need to be excluded and two standard fonts for URLS will need to be agreed (fixed width and variable width).
To my mind, this problem is trivial conceptually, the only complications are as outlined above.
| 1:38 pm on Oct 27, 2009 (gmt 0)|
character != glyph.
e may looks similar to é, but they are different characters. Next y'all will be arguing about whether I can register a domain in Comic Sans or Verdana...
| 2:11 pm on Oct 27, 2009 (gmt 0)|
|"This is the biggest change technically to the internet since it was invented 40 years ago," Peter Dengate Thrush, chairman of the ICANN board, said |
Biggest change to [the] [I]nternet or to computer networking?
When will they stop lying and rewrite history? The Internet was not invented 40 years ago. Not every connection between 2 computer is 'the Internet'! It is just a computer network. And 40 years ago there were no web browsers, no web servers, no hypertext, and certainly no domain names! Even the concept for those was not existing.
The WWW (aka 'the Internet') was invented 20 years ago at CERN by Tim Berners-Lee in 1989, and further refined by him and Robert Cailliau in 1990.
| 2:59 pm on Oct 27, 2009 (gmt 0)|
>> When will they stop lying and rewrite history
That's a very common misconception, perpetrated by mainstream media as they quote each other recursively without checking facts. But to hear it from chairman of the ICANN board, well that's just bad. It sounds like "executive talk" to me, which is a special dialect of hyperbole spoken by people in expensive suits.
But Peter calls it the "biggest change technically". Maybe that's true. I don't really know the magnitude of upheaval it took to make it so.
Underneath this, it's still just a string being converted into an IP address, right? So they are switching to UTF or Unicode or whatever. I'm not saying that it's a simple change, but it's not a paradigm shifting event.
| 3:52 pm on Oct 27, 2009 (gmt 0)|
Imagine just the work that will be needed to recode form validation for domain name and email input...
| 4:06 pm on Oct 27, 2009 (gmt 0)|
|That's a very common misconception, perpetrated by mainstream media as they quote each other recursively without checking facts. But to hear it from chairman of the ICANN board, well that's just bad. |
I came across quite a lot similar statements from senior US official recently...
Well, Joseph Goebbels primary rules were:
|never allow the public to cool off; never admit a fault or wrong; never concede that there may be some good in your enemy; never leave room for alternatives; never accept blame; concentrate on one enemy at a time and blame him for everything that goes wrong; people will believe a big lie sooner than a little one; and if you repeat it frequently enough people will sooner or later believe it. |
| 4:49 pm on Oct 27, 2009 (gmt 0)|
>> recode form validation
oh crap. You're right - I just realized I have a whole bunch of regular expressions out there that are going to fail. It's going to be like Y2K all over again.
only this time I'm not going to dig a bunker in my back yard or brave the WalMart riots for diesel generators and ramen noodles
| 12:32 am on Oct 28, 2009 (gmt 0)|
|e may looks similar to é, but they are different characters. |
Of course they are different characters, but assuming that we don't want phishers and domain squatters, etc. to celebrate this change, they must be treated as identical for the purpose of domain registration - this is an absolute no-brainer (just like case-insensitivity).
| 8:52 am on Oct 28, 2009 (gmt 0)|
I feel really confused.
Could someone try to separate fact from rumor/opinion?
| 9:30 pm on Oct 29, 2009 (gmt 0)|
It is about the TLD part, what is on the right side of the dot.
Domains in Chinese,Russian,Japanese,Hindi are existing since 2000 and are mixed, The domain is in the local language but the part right of the dot is in English, now once the announced changes are implemented (scheduled for 3rd quarter 2010) it will be 100% in the native language.
The registrant (domain owner) that registered/registers example.com in Japanese (com in English), Is the only one that will get the variant of his domain example.com in Japanese where the .com is also in Japanese.
| 8:37 am on Oct 30, 2009 (gmt 0)|
The following 2 messages were cut out to new thread by engine. New thread at: domain_names/4016194.htm [webmasterworld.com]
10:36 am on Oct. 30, 2009 (utc +1)