One of many universal character encoding systems, Rosetta has a home on the web <>, but no body of implementations or users, unlike Unicode or ISO-2022. It defines a character encoding scheme, based around the idea of defining code pages for each language, but so far (the site was last modified in 1997) it only defines characters for English and Russian and a small body of symbols.

There are three goals that stand out upon reading the proposed RFC for Rosetta. The first goal is keeping text size down to the level of legacy character sets. Current Russian character sets, like KOI8-R, use 1 byte per character; Unicode, in its UTF-8 and UTF-16 encoding forms, uses two. The rarely used SCSU Unicode encoding form will compress most alphabetic text to about one byte per character. Secondly, Rosetta tries to keep characters in sort order; this goal is ultimately futile, as many languages count multiple letters as one for sorting (Spanish, Maltese), count one letter as multiple letters (German), have no standard sort order (Chinese, Japanese), or have an otherwise complex sort order (French). Thirdly, Rosetta automatically tags the languages of the document.

Judging purely from an English perspective, Rosetta currently fails to provide all the necessary characters, currently missing the c-cedilla (used in fac,ade), the o-circumflex (used in ro^le), the o-diaresis (used in coo"perate) and an opening quotation mark. These are all easily rectifiable, but shows a lack of research.

It's hard to judge Rosetta on a global sense, besides pointing out that it is currently missing most of the languages and most of the characters of the world. Rosetta does proscribe visual directionality for bidirectional scripts, which does not make Hebrew and Arabic speakers happy.