Verifying numerical algorithms for all of their intended uses is a daunting (impossible?) task. There is a vast middle ground between a code developer who checks a new feature (routine) for a few special cases, and the user, who believing the new feature has been verified, uses the algorithm in a large scale application, which exercises many (untested) possible paths and combinations of parameters. Because the user rarely knows in advance the correct result for the problem being solved, the user is without the ability to identify any algorithmic or coding errors that may exist in the vast amount of numerical results. The Method of Manufactured Solutions (MMS) attempts to bridge this wide gap between verifying a few special cases and fully exercising a routine with the aim of identifying the existence of minor algorithmic or coding errors.