Developers must revise current approaches to standards for analysis software to account for its unique characteristics. When developers think of software standards, they envision software for aircraft cockpits, nuclear reactors, high-speed trains, medical devices, and spacecraft: software that directly controls some component related to the product's safe operation, with or without a person in the loop. Standards for such safety-critical software that have evolved over the past two decades make specific assumptions about the software's development environment. If the software provides computational data that is wrong, and the resulting decision based on that data is faulty, safety is compromised. The software itself is not safety-critical. Nevertheless, safety is involved, sometimes with a lengthy time component and many people in the loop. As with safety-critical software, certain standards address the production of quality computational software or, as it is often called, analysis software. Standards for developing analysis software are based on standards used for other software types. However, the characteristics of analysis software are very different.