I’ve often found truth in the famous George Santayana quote, “Those that cannot remember the past are doomed to repeat it.” That’s an apt warning for what is currently happening—again—with the hundreds of new generic Top Level Domains (gTLDs) that are launching ... and failing to work as expected on the Internet.

First, a quick refresher: As most CircleID readers know, in the early 2000s, seven new gTLDs were launched: .AERO, .BIZ, .COOP, .INFO, .MUSEUM, .NAME and .PRO. Aside from Country Code TLDs (ccTLDs), these were the first top-level changes to the DNS since the early days of the Internet.

Any TLDs that were more than three characters long promptly ran into usability issues. I know this from first-hand experience with .INFO, for which my company, Afilias, is the registry operator. With .INFO’s position as the most popular of all new gTLDs, I spent a good part of my time, in the first five years after .INFO launched, working with vendors to get their systems to accept .INFO email addresses and .INFO domain names as valid. Now, 13 years later, it’s still possible to find systems that reject .INFO addresses. From that experience, I developed my three rules of TLD acceptance.

Mohan’s Three Rules of TLD Acceptance:

1. An old TLD will be accepted more often than a new TLD.
2. An ASCII-only TLD will be accepted more than an IDN TLD.
3. A two or three letter TLD will be accepted more often than a longer ccTLD or gTLD.

Web browsers use different algorithms to improve security and certificate processing. They also have varying rules for how to deal with a website address in a top-level domain that the browser does not recognize (including the use of a technique called search-list processing, which sometimes exacerbates the problem). The use of different algorithms combined with search-list processing can pierce the boundary between private and public administration of namespaces. In addition, many applications and apps that use the Internet still refer to a locally held (and quickly outdated) list of “valid” TLD names, rather than using the DNS to determine domain name validity, despite being warned away from this approach in the IETF’s RFC 3696.

While the issue of universal acceptance never really got solved, the topic takes on heightened importance due to the creation of hundreds of new top-level domains on the Internet. In the earlier set of new TLDs, the primary problem occurred for TLD strings with a length of more than three characters. However, in the current crop of new TLDs, even three character strings get caught in the mix. What was previously considered primarily an infrastructure-level issue is now poised to become a major user-level issue, with negative impact on both the regular Internet user and inside corporations.

In 2003 ICANN’s Security and Stability Advisory Committee (SSAC), of which I am a member, studied problems with the support of new top-level domains by infrastructure and software providers and made several recommendations, including:

• Internet infrastructure providers that have their own customised software for Internet service provision should test the capability of the software to support new TLDs, and correct problems quickly where they are found.
• Internet software application developers should be encouraged to review their software for support of new TLDs. Where problems are found, application developers should upgrade their software, and provide these updates to their user base.
• A central repository of known, commonly used software that has compatibility problems (e.g., DNS resolver software used by common operating systems) with new TLDs, and instructions for how to upgrade the software should be created. This repository would facilitate Internet infrastructure providers and software application developers to provide necessary software updates to users of the Internet to resolve known compatibility issues.

That was over a decade ago! It’s somewhat astonishing that these recommendations are as valid today as they were then, and that readiness is still not measurable.

For example, try to resolve a new TLD such as .email or .onl using a mobile Android device. On my Android phone, typing in www.nic.onl is rejected by the browser and handed off to be treated as a search query, not a DNS query. However, if you typed in the same web addresses on a browser—Chrome, Internet Explorer, Firefox or Safari—in a desktop environment, they go to the right destination, The experience varies even further if you try to resolve a multilingual (Internationalized Domain Name or IDN) top level domain on your browser or email system—and over 100 internationalized top level domains are being activated on the Internet this year. For instance, typing in www.nic.?? or www.nic.xn—6frz82g, the newly delegated .MOBI TLD in Chinese, Safari shows the ASCII string only in the address bar, while Chrome and Firefox show the Chinese string; Internet Explorer dumps it to a search result screen, Users deserve to be taken to the correct destination when they type in an address in their browsers, regardless of the type of device they use or language they type the address in.

The problem isn’t technical in nature; we know how to address the technical issues. What’s needed is coordination and collaboration between far removed actors in the Internet world, so that the software and systems they create can act in a standard way and return a predictable experience to users. The era of hundreds of new TLDs requires new energy, focus and cooperation. New gTLD owners, software developers, network providers and infrastructure companies must work together to ensure their software, browsers, forms, apps, email and other systems are compliant and can handle all delegated new TLDs.

The stakes are even higher than they were in 2001, when I first encountered serious problems with the universal acceptance of .INFO, followed thereafter by .AERO and .MOBI. The need for a coordinated response and clearly visible results to the universal acceptance challenge has never been greater; not doing so could deal a devastating blow to the utility, relevance and legitimacy of all new top level domains. Ensuring universal acceptance should become one of the foremost priorities of all entities engaged in the Domain Name System and using the Internet.