As far as I remember the discussions at Santa Fe, the group wanted to
have its own definition of character set.
It was never spelled out in the minutes or anything like that, but
went something like:
A set of rules for the interpretation of an octet stream, such that:
- The interpretation of each byte cannot be questioned
- The number of representable characters is limited
- No further parameters need to be parsed to get the complete
identity of the character set
The first one means that "UTF-2" probably would be a character set,
but that "ISO 10646" (no encoding specified) would not be.
The second one led to rejection of "ISO-2022" (any octet stream that
uses registered ISO character sets, but doesn't tell you which one)
as a character set name.
The third one led to rejection of Keld's "charset=mnemonic;char-esc=29"
scheme, and he adopted the "mnemonic+29" syntax in his current RFC instead.
In this instance, it would lead to rejecting
"charset=iso10646;char-enc=utf-2" - just in case anyone was thinking of it!
That is my memory of the Santa Fe meeting....
Harald A