The JS 'Decimal' object is patterned after the Java 'BigDecimal' object.
Here is an script:

   a = new Decimal("1.2");
   b = new Decimal("1.1");
   print(a.subtract(b));

Additionally, there are a number of static methods defined based on IEEE
754r standard and the decNumber package.  Here is a script:

   a = new Decimal("1.2");
   b = new Decimal("1.1");
   print(Decimal.subtract(a, b));

Any place a decimal value is expected, any primitive or object may be
substituted.  Doing so will have the effect of constructing a new Decimal
with based on the string value of the primitive or object.  For example,
the following produce the same results as the above:

   print(Decimal.substract("1.2", "1.1"));
   print(Decimal.substract(1.2, 1.1));

In the Java BigDecimal package there are parameters defined as "mc".  Similarly
in the decNumber package, there are parameters defined as "set".  In both
cases, this is expecting a simple object/hash with a round property.  No
other properties are supported at this time.  Example:

  print(Decimal.divide(2, 3, {round:Decimal.ROUND_DOWN}))

Not all exceptions raised are catchable at this time.  What exceptions should
be raised and when is open for discussion.

References:
  http://java.sun.com/j2se/1.5.0/docs/api/java/math/BigDecimal.html
  http://www2.hursley.ibm.com/decimal/dnfloat.html

Sources:
  http://code.intertwingly.net/public/hg/js-decimal/

These Java (and proposed ECMA-262) methods are supported:

 *     abs
 *     add
 *     compareTo
 *     divide
 *     divideToIntegralValue, 2qc
 *     doubleValue
 *     equals
 *     intValue
 *     intValueExact
 *     max
 *     min
 *     movePointLeft
 *     movePointRight
 *     multiply
 *     negate
 *     plus
 *     pow
 *     precision
 *     remainder
 *     round
 *     scale
 *     scaleByPowerOfTen
 *     setScale
 *     signum
 *     stripTrailingZeros
 *     subtract
 *     toEngineeringString
 *     toPlainString
 *     toString
 *     ulp
 *     unscaledValue
 *     valueOf

These Java (and proposed ECMA-262) fields are supported:

 *     ONE
 *     ROUND_CEILING
 *     ROUND_DOWN
 *     ROUND_FLOOR
 *     ROUND_HALF_DOWN
 *     ROUND_HALF_EVEN
 *     ROUND_HALF_UP
 *     ROUND_UP
 *     TEN
 *     ZERO

These Java methods are not supported:

 *     byteValueExact
 *     divideAndRemainder
 *     floatValue
 *     hashCode
 *     longValue
 *     longValueExact
 *     shortValueExtract
 *     toBigInteger
 *     toBigIntegerExact
 
These Java fields are not supported:

 *     ROUND_UNECESSARY

These decNumber (and IEEE 754r) static methods are supported:

 *     abs
 *     add
 *     and (as digitAnd)
 *     canonical
 *     compare
 *     compareSignal
 *     compareTotal
 *     compareTotalMag
 *     copy
 *     copyAbs
 *     copyNegate
 *     copySign
 *     digits
 *     divide
 *     divideInteger
 *     FMA
 *     getExponent
 *     invert (as digitInvert)
 *     isCanonical
 *     isFinite
 *     isInfinite
 *     isInteger
 *     isNaN
 *     isNormal
 *     isSignaling
 *     isSignalling
 *     isSigned
 *     isSubnormal
 *     isZero
 *     logB
 *     max
 *     maxMag
 *     min
 *     minMag
 *     minus
 *     multiply
 *     nextMinus
 *     nextPlus
 *     nextToward
 *     or (as digitOr)
 *     plus
 *     quantize
 *     radix
 *     reduce
 *     remainder
 *     remainderNear
 *     rotate
 *     sameQuantum
 *     scaleB
 *     setExponent
 *     shift
 *     subtract
 *     toEngString
 *     toInt32
 *     toInt32Exact
 *     toIntegralExact
 *     toIntegralValue
 *     toString
 *     toUInt32
 *     toUInt32Exact
 *     xor (as digitXor)

These decNumber methods are not supported:

 *     class
 *     classString
 *     fromBCD
 *     fromInt32
 *     fromNumber
 *     fromPacked
 *     fromPackedChecked
 *     fromString
 *     fromUInt32
 *     fromWide
 *     getCoefficient
 *     setCoefficient
 *     show
 *     toBCD
 *     toNumber
 *     toPacked
 *     toWider
 *     version
 *     zero