Advancing Universal Acceptance of all Domain Names and Email Addresses

ICANN conducted tests by submitting local-language email addresses on websites in various countries, including China and India, to measure how many would accept them. The results revealed that on average only 11 to 12 per cent of websites accepted email addresses in the local language of that country or territory. This demonstrates a substantial gap between the availability of the technology and its actual adoption by downstream applications.

ICANN tested local-language email addresses on 50 Chinese websites across government, media, academia, and business sectors by submitting them on contact pages, and conducted similar tests in India using Hindi email addresses . The results showed that on average only 11 to 12 per cent of websites accepted email addresses in the languages of that particular country or territory .

The technical standards enabling domain names and email addresses in local languages have been developed and implemented at the infrastructure level. However, the software applications that end users interact with — such as websites and mobile apps — have not yet been updated to accept and process these new formats. This gap between standards availability and downstream implementation remains the central challenge for universal acceptance.

Sarmad Hussain noted that the standards are in place and it is now possible to have domain names and email addresses in local languages , but that software applications downstream such as websites and mobile apps still need to be updated to import and process these new local-language domain names and email addresses .

While much of the discussion around UA focuses on the technical supply side, governments represent a powerful demand-side actor. By mandating UA support in the e-services they provide to citizens, governments can generate sufficient demand to make the business case for technology companies to invest in UA implementation. This approach directly addresses the chicken-and-egg problem identified by platform representatives.

Sarmad Hussain explicitly highlighted that governments, as stakeholders on the demand side, can require UA support in the e-services they provide to citizens, and that this would create the demand needed to promote implementation and address the business case challenge raised by Sean Bedford .

Beyond the Unicode layer, the domain name system itself has an additional layer of security through label generation rules (LGRs) developed in collaboration with community panels. These rules are specifically designed to prevent security issues arising from visually similar or confusable characters in complex scripts, including Bangla. ICANN invites stakeholders to review these rules and flag any remaining gaps.

Sarmad Hussain noted that ICANN works with the New Brahmi Generation Panel on many scripts including Bangla, and that rules have been added to address security problems such as visually similar characters when generating strings for the domain name system . He invited the questioner to review the existing rules and raise any remaining gaps .

ICANN has institutionalised its commitment to universal acceptance by embedding it within its five-year strategic plan for FY2026–2030. This plan was informed by an expert working group that drew on input from experts across different stakeholder groups. The inclusion of UA in the strategic plan ensures that resources and attention will be dedicated to addressing these issues over the medium term.

Sarmad Hussain stated that ICANN is concluding an expert working group on universal acceptance that drew input from experts across different states and provided guidelines on how to move the work forward . He confirmed that this work is already part of ICANN's FY26-30 strategic plan, meaning ICANN will be working on these issues for at least the next four to five years .

The failure to achieve universal acceptance is not a niche technical concern but one that affects approximately half the world's population. These 3.5 billion people use languages written in scripts other than Latin and are therefore unable to fully engage with digital services that do not support their scripts. This scale also has direct implications for the business case, as it represents a significant pool of potential consumers.

Guilherme Canela de Souza Godoi stated that around 3.5 billion people use languages written in scripts other than Latin, characterising this as not a marginal issue . He further noted that this represents a significant number of citizens and potential consumers who are not fully engaged precisely because of failures in universal acceptance .

Having internationalised domain names technically available does not fulfil the promise of a multilingual Internet if those names are not accepted by the applications people use daily. When valid domain names or email addresses are rejected because of their script or format, the consequences are concrete: people may be unable to register for services, access education, conduct business, or use public digital services. Guilherme framed this as a human rights issue connected to UNESCO's mandate.

Guilherme Canela de Souza Godoi argued that universal acceptance is about usability and functionality, not just visibility, and that rejection of valid IDN email addresses can affect whether a person can register for an online service, whether a student can access a learning platform, whether a small business can operate using a local-language domain name, and whether a public digital service is genuinely accessible to all citizens . He explicitly characterised this as a human rights issue connected to UNESCO's mandate on multilingualism, freedom of expression, and inclusive digital transformation .

UNESCO has incorporated universal acceptance into its formal monitoring mechanisms by including specific questions on UA in the sixth monitoring cycle of its 2003 Recommendation concerning the promotion and use of multilingualism and universal access to cyberspace. This creates a structured accountability mechanism through which member states must report on their progress. The resulting evidence base can then drive real policy change at the national level.

Guilherme Canela de Souza Godoi announced that UNESCO is launching the sixth monitoring cycle of the 2003 Recommendation and that specific questions regarding universal acceptance have been included in a clear manner . He invited participants to contribute to this monitoring cycle so that their countries' and organisations' experiences with UA are reflected and the evidence can drive real policy change .

Several high-level international agreements already commit all UN member states to advancing multilingualism in digital systems, providing a normative framework for holding governments accountable. The Global Digital Compact and the WSIS+20 outcome document both address multilingualism and universal acceptance, and UNESCO's 2003 Recommendation provides a further legal instrument. These commitments should be translated into concrete national policies and, where necessary, regulations.

Guilherme Canela de Souza Godoi noted that the WSIS Plus 20 outcome document acknowledges progress on IDNs and urges stakeholders to advance their universal acceptance, while the Global Digital Compact calls for multilingualism throughout the life cycle of digital systems . He emphasised that both documents were agreed by all United Nations member states and encouraged the community to hold governments accountable for implementing these commitments . He also referenced UNESCO's 2003 Recommendation as a legal document, albeit non-binding, that countries have agreed to implement .

The Unicode Consortium established a common worldwide standard beginning in 1991 that assigns a unique number to every character in every writing system, enabling computers to represent and process text in all languages. All internationalisation of technology, including the support for complex scripts such as Arabic and Devanagari, builds upon this foundation. Without Unicode enablement, it is not possible for applications to support internationalised domain names and email addresses.

Cathy Wissink explained that all internationalisation of technology starts with Unicode, which began in 1991 by establishing a common standard where every letter of every writing system was assigned a number . She noted that the standards for how complex scripts such as Arabic and Devanagari are displayed, stored, and transmitted through the Internet are primarily done by Unicode , and that it is very hard to think about IDNs and internationalised email addresses without first ensuring technology is Unicode enabled .

Universal acceptance relies on a set of interlocking technical standards produced by three distinct standards development organisations. The IETF, the W3C, and the Unicode Consortium each contribute essential specifications that together enable the internationalisation of the Internet. Understanding the role of these organisations is important for stakeholders seeking to engage with or support the standards ecosystem.

Cathy Wissink identified three standards development organisations involved in the universal acceptance space: the IETF (Internet Engineering Task Force), the W3C (World Wide Web Consortium), and the Unicode Consortium .

Standards development organisations face a persistent challenge in that publishing a specification does not guarantee its adoption by implementers. Developers may be working from older versions of a standard, missing features introduced in later releases, or may not have adopted a standard at all. At the same time, standards must continue to evolve as languages, currencies, and input methods change, requiring ongoing maintenance work.

Cathy Wissink noted that the challenges in the UA space are less about the standards themselves and more about implementation . She gave the example that some implementers may have adopted Unicode 10 while the standard is now at Unicode 17, meaning certain features have not been picked up . She also noted that as countries change their currency formats and symbols, and as new input methods are developed, standards need to be updated continuously .

The work of standards development organisations is largely invisible to end users, who benefit from standards without recognising the organisations and individuals behind them. Yet this work requires continuous investment as languages evolve, new technologies emerge, and standards must be updated. Cathy Wissink called on stakeholders to get involved and to direct resources towards internationalisation standards organisations.

Cathy Wissink described standards development work as effectively invisible to users, who do not recognise the people and organisations behind it . She cited the example of Saudi Arabia recently changing its currency symbol as an instance of the kind of ongoing update work required . She explicitly encouraged people to get involved and to consider adding resources to the space, noting that the world keeps moving and standards must keep pace .

The Unicode standard includes text normalisation processes specifically designed to handle characters that may appear visually similar or confusable, thereby mitigating the risk of phishing and other security attacks. These processes apply not only to Bangla but to many other scripts, including Latin. Implementers should verify that the characters in question are correctly handled through these normalisation protocols, and report any gaps to the Unicode Consortium.

Cathy Wissink acknowledged that confusable characters are a serious concern that the Unicode Consortium takes seriously, noting that such issues exist across many scripts, not only Bangla . She pointed to text normalisation processes within the Unicode standard as the mechanism for addressing confusables , and invited the questioner to check the relevant technical report and contact her directly at Wissink at Unicode.org if the character is not correctly handled through normalisation .

Meta's platform was designed from its inception by a US West Coast company with English as the default language, even though the vast majority of its 3.5 billion users are not native English speakers. This historical design choice has created significant gaps in the platform's ability to support non-English languages and scripts. Addressing these gaps requires deliberate effort and investment.

Sean Bedford noted that while Meta's mission is connecting the world and the vast majority of its 3.5 billion users are not native English speakers, the platform was built with English in mind from a West Coast US-based company base, resulting in many gaps in the way the platform works .

Major platforms such as iOS, Android, and Windows each have integral libraries that determine whether a given string is recognised as a valid URL or email address. These libraries vary considerably in their support for IDNs and non-Latin scripts, with most performing better for standard top-level domains than for different scripts and encodings. When Meta or other application developers deviate from these core platform specifications to add their own support, they create maintenance burdens and potential user confusion.

Sean Bedford explained that platforms such as iOS, Android, and Windows all have integral libraries that dictate how applications built on top of them function, including what counts as a valid URL or email address . He noted that these platforms tend to work fairly well with top-level domains but significantly less well with different languages, scripts, and encodings . He also stated that deviating from a core platform specification creates challenges, maintenance issues, problems, and confusion for users, which is why Meta tends not to do so .

Meta's sign-up process currently rejects IDN email addresses, a limitation rooted in the way the platform was built in 2006 rather than any current technical constraint. However, because the number of users attempting to sign up with IDN email addresses is extremely low compared to standard addresses, it is difficult to justify the engineering investment required to add this support. This creates a self-reinforcing cycle where lack of support suppresses usage, which in turn fails to generate the data needed to make the business case.

Sean Bedford confirmed that Meta does not support IDN email addresses today and that attempts to sign up on Facebook.com with one will fail . He noted there is no technical reason why Meta cannot support this, but that the company will not invest without a strong business case, and the number of attempts to sign up with IDN email addresses is so low compared to non-IDN addresses that it is very hard to rationalise the investment . He described this as a catch-22 or chicken-and-egg situation .

The very low rate of attempts to use IDN email addresses on platforms like Meta is likely not an accurate reflection of genuine demand. Instead, users who would prefer to use their native language are probably giving up and creating ASCII-based email addresses on free providers because it is easier. This means the data that platforms use to assess demand for IDN support systematically underestimates the true need.

Sean Bedford hypothesised that users are simply giving up on IDN email addresses and going with English or a fallback option on a free email provider . He gave the example of a Hebrew speaker who creates an address such as '42 at gmail.com' because it is easier, meaning Meta cannot see from its data that the user's native tongue is Hebrew . He noted this represents an upstream problem where support does not exist on baseline platforms .

Before technology companies and other stakeholders can be expected to act on universal acceptance, they must first be aware that a gap exists and understand its significance. Public sessions, discussions, and discourse such as the one being held serve the essential function of building this awareness. Sean Bedford identified this as the first of his three recommendations for advancing UA.

Sean Bedford listed raising awareness of the gap as his first recommendation, explicitly citing sessions such as the one being held as exactly the kind of activity needed to talk about these topics .

For large technology companies such as Meta, investment decisions are driven by data and business cases rather than by principle alone. Advancing UA therefore requires identifying and articulating strong business incentives that can justify the engineering investment. Sean Bedford noted that for big tech, a compelling business case makes a significant difference to whether and when a change is made.

Sean Bedford identified finding stronger business incentives as his second recommendation, noting that for big tech, having a good business incentive and use case makes a huge difference to being able to make a change . He acknowledged that while the change is technically possible and will eventually happen, it is a matter of when, and that timing depends on the strength of the business case .

There is a significant disconnect between the communities that care deeply about universal acceptance — such as linguists, civil society, and policy advocates — and the engineers and product managers at technology companies who have the power to implement changes. Bringing these two groups into dialogue is essential for translating advocacy into action. Sean Bedford identified this as his third recommendation.

Sean Bedford described his third recommendation as connecting those who care about UA with those who can do something about it, noting that these two audiences tend to live and operate in very different worlds today . He argued that the more these two groups can be brought together for better conversations, the more the space can be driven forward .

Across typical IT and computing degree programmes, students are taught programming, databases, networks, security, and software development almost exclusively using ASCII-based examples and resources. Handling non-Latin languages is rarely integrated into core courses and is instead treated as a special or advanced topic. This leaves graduates without the foundational skills needed to build or maintain systems that support a multilingual Internet.

Fatema Akbar explained that in a typical IT or computing curriculum, courses on programming, databases, networks, security, and software development are primarily taught focusing on ASCII, with textbooks and reference materials also focused on ASCII . She noted that handling languages is typically a special topic in an advanced course or a side project, not integrated within the main curriculum , partly due to the historical significance of ASCII in computing education and partly because textbooks do not include comprehensive coverage of UA principles .

Rather than creating a standalone course, the American University of Bahrain chose to integrate UA topics into existing courses across the computing curriculum in a phased manner. This involved mapping ICANN's 12 curriculum modules to existing courses, training faculty, and rolling out integration semester by semester. The approach ensures that UA concepts are encountered by students from foundational courses through to advanced subjects.

Fatema Akbar described how the university went through training with ICANN on 12 curriculum modules, modified some to match their courses, and trained faculty to deliver them . She explained that they opted for small-scale upgrades to existing courses rather than a standalone elective, integrating UA topics from introductory programming through to security and web development . She also described creating a mapping table between courses and UA modules and a phased integration plan with semester-by-semester rollout .

Given the density of existing computing courses and the administrative burden of formal curriculum changes, small-scale upgrades offer a practical pathway to embedding UA principles. These can include adding UA concepts as requirements for existing assignments and projects, ensuring students encounter and must engage with multilingual handling even when the core lecture content does not change. This approach allows UA to be woven throughout the curriculum incrementally.

Fatema Akbar gave concrete examples of small-scale upgrades, including requiring students in a fundamental programming course to handle any language for input and output from files, thereby prompting them to learn about UTF-8 . She also described teaching students in a database course how to store Arabic text without corruption, and introducing IDNs when covering the domain name system in a networking course .

Integrating UA into computing curricula faces several practical obstacles. Computing courses are already very dense, with new topics such as AI and sustainability constantly competing for space. Formally changing course syllabi or introducing new courses requires a lengthy administrative process. Additionally, faculty members themselves have typically been trained in ASCII-centric computing and require training before they can effectively teach UA topics.

Fatema Akbar identified the density of existing computing courses as the number one challenge, noting that new topics such as sustainability, green computing, and AI are constantly being added . She also noted that changing course syllabi or offering new courses is a lengthy administrative process , and that faculty may need more training because they themselves were trained mostly on ASCII . She noted that the integration work described had been happening over approximately three years .

In Bangla and other complex scripts, certain characters are represented using multiple Unicode code points rather than a single unique code point, creating the possibility of visual similarity between different character sequences. As internationalised domain names expand to include more scripts, this creates a potential vector for phishing attacks where malicious actors could register domain names that appear identical to legitimate ones. This raises both technical and policy questions about how to safeguard users.

Mohd. Abdullah Konu noted that Bangla includes characters such as Ra, Daishundra, and Antostra that are represented using multiple Unicode code points rather than a single unique character . He argued that this may create visual similarity and confusion in domain names, potentially increasing the risk of phishing attacks as IDNs continue to expand . He asked what technical or policy measures ICANN is considering to address these Unicode-related security risks .