Home / Blogs

IDN and Email: The Harsh Reality

There has been a lot of talk about IDNs here and elsewhere but what does the reality look like for a plain user?

As a test, I randomly choose 28 domains from Alexa's top 100 Sites and tried to create a user account with the email address user@宫殿.com.

The bleak result: only wikipedia.org accepted the IDN - and later on failed to send a verification ping - all other rejected the email address.

IDN domains have standardized since March 2003 (RFC 3490) but as far as I know only one major email provider actually supports IDNs flawlessly…

The tested domains in alphabetic order were: adobe.com, amazon.com, bing.com, blogger.com, cnet.com, cnn.com, conduit.com, craigslist.org, dailymotion.com, deviantart.com, ebay.com, facebook.com, hi5.com, imageshack.us, imdb.com, linkedin.com, live.com, livejournal.com, mediafire.com, megaupload.com, mininova.org, myspace.com, nytimes.com, photobucket.com, twitter.com, wikipedia.org, wordpress.com, youtube.com

By Th. Kühne

Related topics: DNS, Domain Names, Email, Multilinguism, Top-Level Domains

WEEKLY WRAP — Get CircleID's Weekly Summary Report by Email:

Comments

Good idea, wrong test. IDN & Email James Seng  –  Nov 21, 2009 7:28 AM PDT

Good idea, wrong test. IDN & Email is still under standardization

http://www.ietf.org/dyn/wg/charter/eai-charter.html

Even RFC 5336 is still only experimental, not full standard yet.

It's complicated Patrick Vande Walle  –  Nov 23, 2009 3:10 AM PDT

James,

Actually, it is a bit more complicated than that. the EAI WG works on how a SMTP server and POP3/IMAP4 server deals with Unicode, mostly on the left hand side of the e-mail address.

I understand from the original posting that the test was more on how forms on web sites accept and process IDNized e-mail addresses. I had a similar post, related to the fact that web forms validate e-mail addresses according to software routines, and that many of these have not been updated for some time. Just google for "email address validation javascript" and you will see what I mean. My concern was more related to the acceptance of new TLDs, but obviously this is also true at the second level.

I guess the result of the test would have been quite different if Th. Kühne had submitted user@xn--3bt078a.com, rather than user@宫殿.com.

This will be a major issue and showstopper for the wide adoption of IDN TLDs in my opinion.

Patrick, noted the differences.But I disagree it James Seng  –  Nov 23, 2009 3:16 AM PDT

Patrick, noted the differences.

But I disagree it is a show-stopper.

It is a chicken and egg problem. Do programmers of the website fixed their javascript and backend first before we deploy or do we deploy and let them encounter problem so they can be fixed?

The issue here is most people don't even realize they have a problem until they encounter it. Afterall, IDN isn't that widely used in most of the US sites.

This is also somewhat true for a lot of Asian website but I will bet they will get fixed faster, once people started using IDN.

IDN's and Internet Protocols Keith Teare  –  Nov 23, 2009 2:17 PM PDT

I agree with James Seng that the inability of a piece of software to support an IDN should be a show-stopper. However there is a more complex discussion of the relationship between a DNS address and a protocol.

When RealNames introduced country and language namespaces back in 1999 we explicitly focused on the browser address bar as an input mechanism and nationally specific and language specific namespaces as an modus operadi. Our reasons for doing so were multiple:

1. The browser sends country and language settings in its header, allowing us to be intelligent about user-inputs and our responses.
2. The ability to set namespace level rules was important to us. Korean, Chinese, Japanese etc all want to maintain their own rules governing how to behave in various circumstances.
3. We wanted to focus on the http protocol, which is a big win as a first-to-market feature, even though we knew adoption for various other protocols would be slower.

When we partnered with veriSIgn in 2000-2001 over IDN v1, http was certainly the target and the most asked for protocol.

So...that said, I think that as IDNs are rolled out, being smart about disclosing issues and focusing on what works now versus what will work later is a key take-away from our experience.

Best
Keith Teare

Oops Keith Teare  –  Nov 25, 2009 9:26 AM PDT

Typo above. I mean that I agree with James that the inability of a piece of software to support an IDN should NOT be a show stopper.

To post comments, please login or create an account.

Related Blogs

Related News

Topics

Industry Updates – Sponsored Posts

Adrian Kinderis Appointed as Chair of Domain Name Association

Internet Reaches 271 Million Domain Names in the Fourth Quarter of 2013

Why We Decided to Stop Offering Free Accounts

The Future of Chinese Domain Names (a Panel Discussion)

dotStrategy Selects Neustar's Registry Threat Mitigation Services for .BUZZ Registry

Tony Kirsch Announced As Head of Global Consulting of ARI Registry Services

24 Million Home Routers Expose ISPs to Massive DNS-Based DDoS Attacks

Afilias Chairman Appointed to Domain Name Association Board

.BUILD Enters Landrush with Support of ARI Registry Services

Dyn Acquires Managed DNS Provider Nettica

Radix Awards Contracts for .website, .host, .space, and .press to CentralNic plc

Afilias Welcomes "Dot Chinese Online" and "Dot Chinese Website" Top-Level Domains to the Internet

Afilias Welcomes .ONL and .RICH to the Internet

Why Managed DNS Means Secure DNS

Neustar Selected to Continue Administering the .US Domain

The Practicalities of a New Domain Launch

Neustar Launches First Site Under '.Neustar' - Its Own Branded TLD

New Internet Top Level Domain ".One" Goes to Highest Bidder in Historic Auction

Dot Desi Reseller, LLC Selects LogicBoxes' Vertical Integration Solutions for the .desi gTLD

Afilias Announces Its First Internationalized TLD With the Launch of .MOBILE in Chinese

Sponsored Topics