Re: Fixed-Point Decimal Proposal

From: p_michel_andre (michel.andre_at_[hidden])
Date: 2002-03-10 10:17:08

Some more thougths?

Maybe the name should be changed to fixed_decimal. 

Since if we later want to include a decimal class with floating 
decimal point (or arbitrary precision) and decimal (radix 10 rounding 
rules and radix 10 exponent), i think that that class should be 
called decimal. 

This is of course not a big issue, but I think it should be 
considered.

I can imagine at least 3 decimal type classes:
1 fixed - the one you propse
2 floating point - like .NETs decimal see 
http://msdn.microsoft.com/library/enzus/cpref/html/frlrfSystemDecimalC
lassTopic.asp (this would be a very useful addition to C++)
arbitrary precision - http://www2.hursley.ibm.com/decimal/ like Javas 
3 BigDecimal or the decNumber C package

/Michel