[rfc-i] Use of unicode
masinter at adobe.com
Mon Apr 23 03:14:30 PDT 2012
Authors' names (ID/RFC author, author of cited document)
EAI Authors' email addresses
Titles of cited documents
IRIs for cited documents
Examples for protocols about I18N and non-ASCII characters
I think these have different priority and different impact. The problems are quite similar to the problems of URIs vs. IRIs: there is a tradeoff between global transcription (everyone can copy text from one device to another, visually, even if haltingly) and local optimization (people get the names and characters they are familiar with).
Avoid spurious uses, like upside-down letters written for fun, dingbats, emoji, etc.
Only use non-ASCII when absolutely necessary; preferably provide ASCII alternative also.
If characters are used for which fonts are not universally available, release at least one format with embedded fonts (PDF/A, ePub).
More information about the rfc-interest