Our font icon use common letters. We already know this a problem for accessibility and Google Translate (see #11)
Good thing, new icons like those use for calendars, use the Unicode private range.
But we are still using daily on new pages the old icons which use letters. We know we are increasing the problem and our future technical debt.
We should duplicate all these icons to the private range and update the docs with corresponding codes. All new development would use these new codes, but legacy code would still work.
You’re right, I’m using these new guidelines for the new icons added to the font.
I now use the private area of UTF-8 or the reserved slots in UTF-8 characters. For example to add a “print” icon I search this keyword here: http://www.fileformat.info/info/unicode/char/search.htm
I get these results: http://www.fileformat.info/info/unicode/char/search.htm?q=print&preview=entity
The first char is ok to do the job: http://www.fileformat.info/info/unicode/char/2399/browsertest.htm
So I will use this slot. This pages shows the version of Unicode standardisation process, and all the code values (HTML, CSS entities): http://www.fileformat.info/info/unicode/char/2399/index.htm
I prefer using the UTF-8 slots whenever it’s possible. It gives an added semantic value to the characters. Otherwise I use the private area.