The latest Java toy to be released to developers through IBM’s AlphaWorks site is the BigDecimal class. According to the company, BigDecimal provides Sun’s Java programming language with the floating decimal point it currently lacks. Most computers use only binary floating point arithmetic. While it can’t precisely represent common decimal values like 0.1, binary floating point is widely used where performance is the main concern. IBM argues that for most Java applications, performance is no longer a significant issue. Using decimal floating point in those applications avoids anomalies and yields results that match human expectations. This should make it easier to add human-oriented arithmetic to applications written in Java. The new class implements immutable arbitrary-precision decimal numbers with both scientific and engineering exponential notations available as options. By distributing the class to developers for free, IBM hopes to evaluate its utility and functionality.
