Principles

The goal: predictable sound from spelling

English orthography records history, not pronunciation. The consistency English started with was lost over several centuries, and that is what Euspell aims to reverse: not strict phonetics, but predictable meaning within the existing spelling convention, to improve both human and machine comprehension. The only new forms are qh and uw, easy to get used to quickly.

The constraints

A phonetic ideal, pursued naively, produces something unreadable and politically dead on arrival. Euspell deliberately trades some phonetic purity for legibility and adoptability:

• Change only what earns it. The majority of words are left exactly as they are (encoding 000). Reform is spent on the words whose spelling actively misleads, not on churning the whole language.
• Preserve the silhouette. Reformed words keep as much of their original shape as possible — night → niht, not a wholesale respelling — so fluent readers adapt in minutes, not months.
• One grapheme, one reading. Removing exceptions has high priority: the sounds of graphemes such as ea, ch, ou and ow will be highly consistent and predictable. Where traditional spelling collapses distinct words into one form, the reform splits them — records (noun) vs recordz (verb) with unambiguous meanings.
• Reversibility. Conversion round-trips: revertText(convertText(x)) === x, so reformed text can always be turned back into traditional English.

The hard problem Euspell commits to solving

Reform proposals usually stall on homographs — words spelled alike but pronounced differently, like read (present) and read (past), or the noun/verb pair record. A context-free respelling cannot tell them apart. Euspell treats this as the central engineering task rather than an edge case: a part-of-speech tagger and a set of disambiguation rules decide, word by word, which reformed spelling applies.