Traditional Computing against Soft Computing
Soft computing actually is approximate computing, the reason we using soft computing is that we can not understand the nature so clearly. But in artificial intelligence field the reason is more relative robust of system. No matter randomness or fuzziness, traditional computing will against them.
Example 1,
In rock orbit we need accuracy and precise computing, say traditional computing, if not, a rock send and run in the orbit will lost in endless universe.
Example 2,
In a robot domain, such as, a robot press the button of a lift, we do not need to calculate so accuracy whereas only press the button is enough. Remember that, the button is round and not a point.(There are many human knowledges are uncertainty knowldeges)
From that two example, many application so accuracy or precise is not necessary. Using traditional computing or soft computing is depending on what application you doing. Between traditional and soft is the issue of accuracy/precise.