民用计算工具中算术运算误差的掩盖法

Glossing over Error of Arithmetic Calculation in Design of Civil Calculating Tool Software

给出了一种根据输入数据的十进制特征判断算术运算结果准确值的最低非０位并据以避免误差的掩盖方法，用于民用计算工具软件的设计，可使较简单的计算中误差完全不出现，同时在复杂的运算中又不降低原有精度．此算法已用于电脑报税器的软件设计中，效果很好．

The method of glossing over error introduced in this paper is: from the decimal features of input data,to deduce the least significant non 0 place of the exact result of arithmetic calculation , and round the approximate value off at this place. This method used in design of civil calculating tool software can make simple calculation free from error entirely, and make users place more trust in reliability of the products. Besides, this method is never bad for the precision of complicated calculation.