I was playing with code pages and ran into an interesting case:  Code Page 21027 - Ext Alpha Lowercase. 

This code page has some interesting behavior.  It looks like a Japaneses EBCDIC code page, however its kind of "missing" mappings for some characters, like 8, 9, =, H, I, Q, R, Y, Z, Halfwidth Katakana Ku, and Halfwidth Katakana Nu.  If you try to convert these characters from Unicode to 21027 and back, you end up with other code points.  For example the Unicode character "8" (U+0038) changes to U+f8f8 when its round tripped in this code page.

If anyone has is using this code page or knows what its used for I'd be interested in hearing about it :-)

In .Net v2.0 we've removed support for this code page, although it was available in v1.0 and v1.1 and continues to be available natively in Windows XP and Server 2003.