My understanding of "full-width ASCII" is to support roman-ji in Japanese, where such niceties as proportional spacing are just now appearing in the public marketplace. These code points (U+FF00 - U+FF60) are supported primarily so that fonts that contain both the "standard" ASCII and the "full-width ASCII" (e.g. Monotype Andale, Arial Unicode, Mincho, etc.) can differentiat which glyph to use.

In looking up the code point range for the full-width ASCII, I also discovered that there really are code points for the various arabic presentation forms, as well as Latin presentation forms and others. So, within UNICODE, you can explicitly define the correct glyph for presentation, even if you have a font engine capable of "fixing it up" for you. Course that does complicate UNICODE even more, as someone who is "using UNICODE" cannot be known to be using the presentation forms, so the rendering engine must be capable of passing through the presentation forms, and "massaging" the un-presentation forms when necessary. Sheesh!

That's why I love IWETHEY...you learn something new even when you don't expect to....