For odd and extinct characters or special alphabets, pending any actual coding by Unicode, it might be more desireable to treat them as a cyphers for the extended Latin alphabet rather than encoding them in the PUA.

For example, if someone is using a character for the _ng_ sound in _singer_ then encode it as eng, lowercase U+014B, uppercase U+014A, regardless of its appearance.

A special font would be needed to view the charactes correctly, but that is also true if the shapes are encoded in the PUA.

The advantage to treating such alphabets as cyphers is that a user with some knowledge of IPA can read material written in them without having to learn the characters, but can also see the original appearance of the text simply by switching a font.

Greek characters and Cyrillic characters might also be used.

Some of the matching might be rather arbitrary.

But there are doubtless thousands of various revisions of the Latin alphabet and invented scripts out there, most of which were scarcely or at all used by any but their inventors (as well as cypher systems like Morse Code and signal flags).

There seems to me little to be gained from Unicode attempting to encode every one of these separately (along with any new systems people might invent.)

Shavian, for example, might well have been encoded as an IPA cypher, simply by providing a cross-reference between Shavian character and Unicode IPA character.

If a large amount of Shavian were to be placed on the web, and I wished to investigate it, I would probably create an IPA type font with appropriate IPA glyphs assigned to Unicode Shavian values, so I could investigate the material more easily without having to learn the Shavian script.

Similarly with some syllabic systems I might create a Latin letter font with Lattin letter digraphs replacing the syllabic characters.

Indeed, a Latin letter only Unicode font covering many non-Latin scripts might be quite useful, if one could get used to reading the words spelled backwards in Hebrew and Arabic, along with superscript and subscript vowels.


Jim Allan






Reply via email to