Gunther wrote:
> My claim was that Unicode is designed to be ISO Latin-1 backwards
> compatible and therefore there should be a Latin-1 backwards
> compatible UTF. This argument has never been defeated by anyone who
> resopnded.
I think what we have here is an example of an enthymeme. The missing
premise is stated explicitly here:
Premise A: Unicode is designed to be ISO Latin-1 backwards compatible.
Premise B: Character sets with which Unicode is backwards compatible
require a UTF.
Conclusion: Therefore, [Unicode requires] a Latin-1 backwards compatible
UTF.
The problem, of course, is that there is no general assent as to the
truth of Premise B. As others have pointed out, Unicode is backwards
compatible with a number of character sets--JIS X 0208, for example.
But it is not self-evidently true or agreed that each such compatible
character set requires a UTF.
If the unstated Premise B is false, then the conclusion itself is false,
and the syllogism is faulty.
I could be more restrictive in the interpretation of what Gunther
means by "backwards compatible"; by reading between the lines, it
appears he doesn't mean cross-mappable or interconvertible, but
instead "encoded so that the entire non-ASCII portion of the
8-bit character set (0x80..0xFF, or 0xA0..0xFF, depending) is
represented in Unicode by a constant offset and data-widening to
16-bits". In this case, there is exactly and only one character
set which meets this requirement: ISO/IEC 8859-1 Latin-1. Calling
this concept "offset compatibility" and restating the syllogism
in this more restrictive form, we have:
Premise A: Unicode is designed so that Latin-1 is offset compatible.
Premise B: Character sets that are offset compatible to Unicode
require a UTF.
Conclusion: Therefore, [Unicode requires] a Latin-1 offset compatible
UTF.
Even in this form, however, if Premise B is not true, the conclusion
fails.
It seems to me that the consensus of the list in this regard is that
Premise B is not true.
Q.E.D.
--Ken Whistler
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:41 EDT