kau'ai KAUhAI experimental cmavo

Microdigit-spanning endianness binary-toggle.

Switches the endianness of the microdigits in all subsequent numeric strings (during their interpretation phase) from big-endian to little-endian or vice-versa; macrodigit endianness is unaffected; it does span across macrodigits (into their microdigits), until used again or terminated by the end of scope of li, etc. (see below). So, for example, in the context of specifying a month of the year and a day of the month (in macrodigit big-endian format: MM-DD), exactly "kau'ai pa re pi'e pa no" is in little-endian mode for exactly the microdigits and means only the first day of the twenty-first month, if it existed, (and not the twelfth day of the tenth month, the tenth day of the twelfth month, the twenty-first day of the first month, or anything else); meanwhile, exactly "kau'ai kau'ai pa re pi'e pa no" is big-endian in its microdigits and means only the tenth day of the twelfth month (and not the twenty-first day of the first month, the twelfth day of the tenth month, the first day of the twenty-first month (if it existed), or anything else). Without prior specification or context, Lojban assumes a big-endian setting (although such language is never explicitly used in the CLL), except for certain brivla (such as detri, currently); in such a case or in any case whatsoever other than having previously and still-actively and explicitly specified the little-endian interpretation, including bi-endian or middle-endian etc./vel sim. cases, the first active and explicit usage of this word (called the "original usage" here) switches the interpretation to little-endian; thereafter, each usage will toggle between big-endian (all odd-numbered subsequent occurrences) and then back to little-endian (all even-numbered subsequent occurrences), where the original usage (which switched away from contextless default or non-little-endian interpretation) is counted as the zeroth usage. This function is an involution (at least after the original usage). It terminates with the closing of the scope of li, the usage of a brivla or gadri or du (or similar) or boi, or the overarching specification for the text (iff used therein); it does span across (not terminate with) VUhU and the like. For now, it really should not be used twice within the same numeric string/macrodigit. Notice that exactly "kau'ai pa re ci" means "three hundred twenty-one" because there is only one macrodigit (so, its remaining big-endian, assuming contextless default, is inconsequential), but its microdigits have become little-endian. See also: kau'au.

In notes:

Macrodigit-spanning endianness binary-toggle.
x1 (digit string/byte, storage system, convention) has endianness x2 ("ce'o" sequence of numbers (li); description (ka?)); x1 is x2-endian.