>Any unassigned code point you choose will get the missing glyph >display for the font that happens to be selected for that character (on >many systems that's no longer easily predictable, due to font >substitutions). In short, there's no guarantee that *any* specific code >point will give you the 'typical' display.
Isn't font substitution more and more becoming the rule? Certainly I expect Mac OS X to search all its fonts for glyphs for a given codepoint, and when none can be found, I get the elegant symbols in the "Last Resort" font, all of which have a similar form, as described by Deborah Goldsmith, namely the glyphs from http://www.unicode.org/charts/ expanded to show the name and Unicode scalar value range of the block in question. This sure seems like a nice way to address the issue.