Dimensional Analysis in C++ ETH Z（量纲分析在c++中乙Z）.pdf

Dimensional Analysis in C++ Scott Meyers, Ph.D. Software Development Consultant smeyers@ Voice: 503/638-6028 / Fax: 503/638-6614 The copyright to this material is held by Scott Meyers Page 1 or by Addison Wesley Longman, Inc. Last Revised: 11/10/02 Dimensional Analysis in C++ Scientific and engineering calculations are dependent on correct use of units in calculations: It makes no sense to assign a time value to a distance variable It makes no sense to compare a mass variable with a charge variable But most software ignores such units: double t; // time - in seconds double a; // acceleration - in meters/second2 double d; // distance - in meters ... cout d/(t*t) - a; // okay, subtracts meters/sec2 cout d/t - a; // should be an error, as it // subtracts meters/sec and // meters/sec2 The copyright to this material is held by Scott Meyers Page 2 or by Addison Wesley Longman, Inc. Dimensional Analysis in C++ Typedefs just disguise the problem: typedef double Acceleration; typedef double Time; typedef double Distance; Time t; Acceleration a; Distance d; ... cout d/t - a; // still compiles, but is still wrong We want a way to use the C++ type system to: Make unit compatibility errors impossible: ➠ They’ll be detected during compilation Do so with minimal runtime performance impact: ➠ Minimal memory overhead, minimal runtime over