For two millennia, the name of the god of light was written Ἀπόλλων. By the time it reached the early internet, it had been flattened to apollo — seven lowercase ASCII characters, no breathing mark, no acute accent, no long vowel. The domain system that runs the web was built on a seven-bit alphabet, and that alphabet could not carry the pitch, the breath, or the duration that ancient Greek used to distinguish one word from another.
The Three Coordinates of a Greek Name
Every classical Greek name carries three phonological coordinates that the Latin alphabet usually ignores:
- Rough breathing. The initial h-sound in Ἀπόλλων and Ἅιδης is not ornamental. It changes the etymology, the meter, and the pronunciation. Dropping it turns a god into a ghost of himself.
- Stress. The acute (´) or circumflex (῀) tells the reader which syllable received the musical pitch accent. Ancient Greek was pitch-accented, not stress-accented; the accent was a rise and fall in tone, not merely volume.
- Quantity. Long vowels — η, ω, long α, ι, υ, and many diphthongs — determined meter, morphology, and even meaning. In the PÚNYCODEX tier system, a name that preserves both stress and length qualifies for Tier 1.
Take Apóllōn. The full restoration Apóllōn preserves the acute on the second syllable and the long omega in the final syllable. Remove either coordinate and the name is still readable, but it is no longer the same philological object.
From Breath to Bit
Unicode solved the character problem: Greek Extended, Combining Diacritical Marks, and the basic Greek blocks give us every breathing, accent, and macron we need. The IDNA2008 standard, in turn, lets those characters travel through the DNS via Punycode. What remained was a curatorial problem: which forms are historically justified?
PÚNYCODEX does not invent accents. The canonical lexicon in type/js/lexicon.js is anchored to reference works such as Liddell–Scott–Jones (A Greek–English Lexicon), the Oxford Classical Dictionary, and Beekes's Etymological Dictionary of Greek. When a name is marked Tier 1, it means the Greek original has both stress and length and the Unicode restoration follows a single, defensible scholarly orthography.
Philological Justice
There is a small but growing movement to treat domain names as cultural identifiers, not just routing labels. A Unicode domain is a chance to give a name back its original spelling — not for novelty, but because the spelling encodes information.
The dual-tier names — Apóllōn, Hekate, and Nike — are the most dramatic example. Each has multiple historically valid Unicode restorations, each corresponds to an owned domain variant, and each variant is a real, attested alternate spelling rather than a modern convenience. They are not typos; they are witnesses.
What This Means for the Web
As internationalized domain names become more common, the question is no longer "Can we write Greek in a URL?" but "Which Greek do we choose?" PÚNYCODEX argues for the most accurate form that is still typeable and ownable. Sometimes that means a fully accented restoration; sometimes it means a macron-only fallback such as Apollōn; sometimes it means accepting a plain ASCII form only as a last resort.
The goal is not purity. The goal is transparency: every temple page explains which coordinates are preserved, which are compromised, and why. If you want to see the engine in action, type Apóllōn in the type tool and watch the ASCII-to-Unicode mapping unfold character by character.