Acceptance of IDNs: The 14th-Century Roots of a 21st-Century Problem

06 October 2025 | By Mark W. Datysgeld

Concerns about the acceptance of IDNs and Universal Acceptance (UA) belong to the Internet era, with the intent of supporting valid domain names and email addresses across all software, systems, and user interfaces. However, the structural dynamics behind supporting varied writing systems long predate standards bodies and software stacks. The core issue of systems being optimized around a compact Latin-based repertoire emerged in the late medieval print economy, and was reinforced at each subsequent communications revolution. IDNs, in that sense, are a continuation of a fourteenth-century problem.

Early printing systems

Movable type as a printing method was developed circa 1040 by Chinese artisan Bi Sheng, who cast blocks in clay which allowed for the reusable and consistent reproduction of characters in print (McMurtrie, 1943). In 1377, Korea produced the oldest book in existence printed with movable metal type, the “Jikji” (Hangul transcription: 백운화상초록불조직지심체요절), confirming that the technique was in working usage before the European breakthrough achieved by Johannes Gutenberg’s printed Bible in 1454-1455 (UNESCO, 2004).

Regardless of which is the oldest document, early usage of movable metal type can be traced back to East Asia, predating the European implementation by at least several decades, or perhaps more than a century. So why is it that the printing revolution took place in Germany instead? One essential factor was the compactness of the Latin alphabet. With a few dozen base letters plus numbers and essential symbols, the existing technology allowed for the mass-production of uniform blocks (Encyclopedia Britannica, 2025). The blocks could be cast in manageable quantities, typesetters could retrieve them rapidly from small cases, and printing could be carried out at greater scale and at lower prices.

While the same capabilities existed in East Asia, the economics diverged because of the writing systems involved, creating a different script-system fit. Chinese publishing frequently preferred woodblock printing for reasons of logistics, storage, reusability, and labor; even with access to movable type. This was due to the thousands of characters required for general literacy, which resulted in vast inventories and slower retrieval (Needham and Tsuen-Hsuin, 1985). In this sense, printing became less practical and scalable, with raised unit costs and slowed throughput.

Evolution and dependency

This script-system fit became a path dependency. When the electrical telegraph’s development advanced enough to become more widely available, kickstarting the next communication revolution, standardization assembled around the simpler Latin repertoire. The 1865 International Telegraph Conference in Paris, which founded the still-existing International Telegraph Union (now the International Telecommunication Union), established Morse as the international telegraph alphabet (ITU, 1865).

Workarounds emerged for non-Latin writing systems, much like today we implement solutions such as Punycode to enable IDNs on the DNS. China’s standardized solution emerged in the 1880s with the numeric Chinese Telegraph Code (CTC), a codebook mapping characters to numbers so that operators could transmit ideographs using Latin-centric channels. This complex standard featured over 7,000 common codes, in four-digit code groups, 0000–9999 (Huang, 2019). This became an efficient layer of mediation, but never became as practical as standard Latin transmission.

Teleprinter standards compounded the effect, with the five-bit code “International Telegraph Alphabet No. 2 (ITA2)” standardized by the ITU and used throughout Telex, which was also engineered for a small set of characters from the Latin repertoire (ITU, 1994). When the “American Standard Code for Information Interchange” (ASCII) was approved in 1963, it did so in continuity with this tradition: a 7-bit code optimized for Latin.

By the time the Internet’s more formal naming standards were written in the 1980s, the Latin-ASCII assumption felt logical and in line with the existing conventions; something that did not need to be challenged, especially in the face of existing computing limitations and needs. RFC 952 (1985) and its clarifications in RFC 1123 (1989) defined hostnames in terms of the ASCII LDH pattern (Letters, Digits, Hyphen), thus preserving a Latin-only label space at the protocol edge (at any level).

When the IETF later internationalized domain names, it did not lift the ASCII constraint, instead, IDNA places non-ASCII labels into the DNS as “A-labels” (Punycode with the “xn--” prefix), while users see and type ‘U-labels’ (Unicode), as per RFC 3492 and RFC 5890. Email is a positive outlier in that the SMTPUTF8 extension (RFC 6531) permits UTF-8 addresses and headers, but adoption remains partial and uneven across the software and infrastructure landscape, which significantly limits usage of this feature.

Conclusion

Seen across seven centuries, this is a continuous dependency chain: a compact alphabet gave European printers a manufacturing advantage; telegraphy and teleprinters generalized that compactness into code; ASCII encoded it into computing; Internet standards etched it into naming and addressing. When the internationalization community insists on IDN adoption and support, it is pushing back against forces deeper than a one-off protocol choice.

As Mullaney (2018) puts it, this rests on a “presumed alphabetic universalism in the form of Morse Code, Braille, stenography, Linotype, punch cards, word processing, and other systems developed with the Latin alphabet in mind.” The fact of the matter is that we as a collective of technologists and policymakers finally stand a chance of breaking away from this pattern. This decision is not just in the hands of small number of engineers and standards developers, but rather in individual decisions we make on how to set up our systems properly.

Procurement should treat UA and EAI as a requirement, not as a nice-to-have.

By Mark W. Datysgeld