Looking at the two sets of Brahmi numbers would also be instructive... Sent from my Android phone On Jul 12, 2012 6:21 AM, "Richard Wordingham" < richard.wording...@ntlworld.com> wrote:
> What is a number having a numeric type of "digit" meant to convey? > > The old Unicode 2.0 definition definition of "digit value" seemed clear: > "Digit value. This is a numeric field. If the character represents a > digit, not necessarily a decimal digit, the value is here. This covers > digits which do not form decimal radix forms, such as the compatibility > superscript digits. This field is informative." > > That definition seems to be gone. From what we now have, I can think > of several meanings, e.g.: > > 1) It's a digit in a system of decimal place notation, but doesn't quite > qualify for some reason. Typical examples are: > > a) U+19DA NEW TAI LUE THAM DIGIT ONE - cruelly denied "decimal" status > because it wasn't assigned with 9 clones of the other Tai Lue digits. > > b) U+2070 SUPERSCRIPT ZERO - not in a contiguous range, and > apparently possibility of misleading parsers > > c) U+2080 SUBSCRIPT ZERO - apparent possibility of misleading parsers > > 2) It's in a decimal system, but not with place notation: > > a) U+10E60 RUMI DIGIT ONE. By contrast U+10E69 RUMI NUMBER > TEN is a mere "numeric" - possibly because numeric field values of > blank for decimal digit value (field 6 in UnicodeData.txt), 1 as the > digit value (field 7) and 10 as the value (field 8) would be too > confusing, as well as contrary to the current rules. > > On the other hand, I don't see why, apart from a general disapproval of > compatibility characters, the Roman numerals U+2170 SMALL ROMAN > NUMERAL ONE to U+2178 SMALL ROMAN NUMERAL NINE don't count as digits. > > 3) It's derived from a decimal digit, e.g. U+2468 CIRCLED DIGIT > NINE is "digit", whereas the next in the series, U+2469 CIRCLED NUMBER > TEN, just has a numeric type of "numeric". > > ---- > > It's not clear to me why the following decimal digits (in the normal, > not the Unicode sense) are not classified as "digit" but just as numeric > > U+1D360 COUNTING ROD UNIT DIGIT ONE > U+3021 HANGZHOU NUMERAL ONE > > The only reason for U+1D369 COUNTING ROD TENS DIGIT ONE not to be a > digit that I can think of is that the system is conceived of as a > centesimal system. The counting rods 'UNIT' and 'TENS' digits are used > alternatively to avoid misreading, with various methods > for indicating zero. > > Likewise, why are U+0C79 TELUGU FRACTION DIGIT ONE FOR ODD POWERS OF > FOUR and related characters not digits? Is it because they are a base > 4 (or collectively hexadecimal) system? > > Perhaps some light can be shed on the system by learning what people > actually use the numeric types and (decimal) digit values for. > > Richard. > >