From these examples, we see that special units in a subfield of physics can be dictated by fundamental theory (as in atoms) or by the size of standard objects, or by a prevailing condition (such as atmospheric pressure), or by experimental practice (sometimes, obsolete practice), or by a combination of good reasons. One can go all the way in this direction and define special units for each problem. For instance, if we discuss a pendulum, we could use its mass as the unit of mass, the distance from the pivot to the center of mass as the unit of length, and also set g=1. Proceeding in this way is equivalent to carrying out a dimensional analysis for the problem at hand, with the final result of compressing a large amount of information into a few graphs of dimensionless functions of dimensionless parameters. Unfortunately, the highly compressed information may become unintelligible, and in any case some time must be spent to decompress it. Maximum concision is for specialists only.