From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Mon Apr 11 2005 - 08:35:35 CST
From: "N. Ganesan" <naa.ganesan@gmail.com>
>> but the proposed new glyph is a vertical oval [for Tamil]
>
> Many months ago, we at Infitt working group (http://www.infitt.org)
> on unicode proposed Tamil digit zero. It already came into effect, Tamil
> digit
> zero exists now in Unicode 4.1:
> http://www.unicode.org/charts/PDF/U0B80.pdf
>
> Similarly, a zero digit glyph of circular form used in Malayalam
> books should be shown in the Malayalam code chart.
I note that in Windows, the Regional settings include a selection for
Malayalam digits (they are part of the default settings for the Malayalam
locale):
================================================
Number format:
- Measure units system: metric
- Standard national digits:
from MALAYALAM DIGIT ZERO..NINE (U+0D60..U+0D69)
- Digit substitution mode:
- "inactive" by default (keep ASCII digits).
- the "national" mode always replaces ASCII digits by national digits.
- the "contextual" mode replaces ASCII digits only after malayalam text.
- Decimal symbol: "."
- Number of decimals: 2
- Integral digits grouped by: ","
- List separator: "," (should probably be manualy edited and replaced by
";")
- Negative sign: "-"
- Negative format: "-൧.൧" (if national substitution active)
or: "-1.1"
- Leading zero: "൦.൭ (if national substitution active)
or: "0.7"
- positive: "൧൨,൩൪,൫൬,൭൮൯.൦൦" (if national substitution active)
or: "12,34,56,789.00",
- negative: "-൧൨,൩൪,൫൬,൭൮൯.൦൦" (if national substitution active)
or: "-12,34,56,789.00"
================================================
Currency format:
- currency unit: "ക" (MALAYALAM LETTER KA and NBSP)
- unit position: unit NBSP sign number
- positive "ക ൧൨,൩൪,൫൬,൭൮൯.൦൦" (if national substitution active)
or "ക 12,34,56,789.00",
- negative "ക -൧൨,൩൪,൫൬,൭൮൯.൦൦" (if national substitution active)
else "ക 12,34,56,789.00",
- same formatting for the numeric value
================================================
Time format:
- hour mode: 00 to 23 (no AM/PM indicator)
- separator: "."
- template: "൧൪.൪൯.൩൦" (if national substitution active)
or "14.49.30"
================================================
Short date format:
- calendar type: gregorian
- separator: "-"
- template: "൧൧-൦൪-൦൫" (if national substitution active)
or "11-04-05"
Long date:
- template: "൧൧ ഏപ്റില് ൨൦൦൫" (if national substitution active)
or "11 ഏപ്റില് 2005"
================================================
So there may already exist LOTS of Malaylam documents created under Windows
(including emails with Outlook Express, Word and Excel documents) showing
the one-half glyph for digit 0, in dates, currency amounts...
This email is composed in Outlook Express with the "contextual" digit
substitution mode, that's why the currency ammount that follows the currency
unit symbol is using Malayalam digit zero (shown currently with the one-half
glyph which is part of the windows XP standard fonts for Malayalam) and not
ASCII digit 0; same thing in the date (see how 2005 is rendered after the
local month name: it shows the horizontal bar with left curl, i.e. the
one-half glyph).
This makes a huge precedent, and probably there are now tons of publications
using this glyph, even if Malayalam mathematic historians would have
objected about it.
So there's now:
- a incorrect but widely used "modern" Malayalam positional decimal system
using the one-half glyph for zero, and no tens digits.
- a correct but rarely used "modern" Malayalam positional decimal system
using ASCII digit 0, or better the existing Malayalam letter TTHA (probably
by a few Malayalam purists that have manually corrected their text documents
and don't use spreadsheets like Excel): these users would benefit from the
change, but they are probably used by purists.
- the historic numbering system without a zero, but with additional digits
for powers of ten multipliers 10, 100, 1000 (possibly also conjuncts of
these multipliers for higher numbers).
Suppose that Malayalam glyph is changed now, will Microsoft update its core
fonts for Malayalam in Windows XP or Office ("Kartika" and "Arial Unicode
MS")? If it does so, what will happen to documents that were initially
written by Malayalam purists that used the one-half glyph to actually mean
the one half value, and used a ASCII DIGIT ZERO or MALAYALAM LETTER TTHA to
unambiguously mean the null value?
If the glyph is changed, the Malayalam number format in Windows (in
"national" mode) will look correctly like this:
positive number: "൧൨,൩൪,൫൬,൭൮൯.ഠഠ" == "12,34,56,789.00"
negative number: "-൧൨,൩൪,൫൬,൭൮൯.ഠഠ" == "-12,34,56,789.00"
positive currency: "ക ൧൨,൩൪,൫൬,൭൮൯.ഠഠ" == "ക 12,34,56,789.00"
negative currency: "ക -൧൨,൩൪,൫൬,൭൮൯.ഠഠ" == "ക -12,34,56,789.00"
time: "൧൪.൪൯.൩ഠ" == "14.49.30"
short date: "൧൧-ഠ൪-ഠ൫" == "11-04-05"
long date: "൧൧ ഏപ്റില് ൨ഠഠ൫" == "11 ഏപ്റില് 2005"
(just above I used LETTER TTHA to show the correct glyph, it loks like a
circle with the "Kartika" font as well as in the "Arial Unicode MS" font),
instead of these "incorrect" formats:
positive number="൧൨,൩൪,൫൬,൭൮൯.൦൦" == "12,34,56,789.00"
negative number="-൧൨,൩൪,൫൬,൭൮൯.൦൦" == "-12,34,56,789.00"
positive currency="ക ൧൨,൩൪,൫൬,൭൮൯.൦൦" == "ക 12,34,56,789.00"
negative currency="ക -൧൨,൩൪,൫൬,൭൮൯.൦൦" == "ക -12,34,56,789.00"
time: "൧൪.൪൯.൩൦" == "14.49.30"
short date: "൧൧-൦൪-൦൫" == "11-04-05"
long date: "൧൧ ഏപ്റില് ൨൦൦൫" == "11 ഏപ്റില് 2005"
So may be the solution for Microsoft in Windows will be to not change the
glyph at all in its fonts, and instead display the glyph of LETTER TTHA when
in "national" digits substitution mode.
For historic traditional texts, there's the need to define characters for
10, 100, 1000, symbolized below by "X", "C", "M" (Roman digits) in this
example:
"൧M൨C൩X൪Y൫X൬M൭C൮X൯"== "12,34,56,789"
(probably there's another digit with a definite name for the 100,000 value
noted "Y" here, according to the number spelling that groups every two
digits above thousands, as noted in the numeric formats above; I can't
represent decimals above as it would require using fractions).
The problem may probably extend as well on MacOSX (I don't know if MacOSX
has a Malayalam locale)...
This archive was generated by hypermail 2.1.5 : Mon Apr 11 2005 - 08:36:54 CST