In musical temperament, intervals are usually stated in <i>cents</i>, i.e.&nbsp;hundredths of a 12-EDO semitone. A more convenient unit would be 1/12276 of an octave, i.e.&nbsp;1/1023 of a 12-EDO semitone. I&nbsp;call this a <i>mil</i> because it is nearly 1/1000 of a semitone (cf.&nbsp;1KB&#8239;=&#8239;1024&#8239;bytes). Thus the frequency ratio <i>f</i><sub>2</sub>/<i>f</i><sub>1</sub> is expressed in mils as 12276&#8239;log<sub>2</sub>(<i>f</i><sub>2</sub>/<i>f</i><sub>1</sub>). Advantages of the mil include: (i)&nbsp;its size is almost self-explanatory; (ii)&nbsp;rounding to the nearest mil is accurate enough for practical purposes; (iii)&nbsp;<em><b>numerous important intervals are very close to whole numbers of mils</b>:</em>

| Interval | mils
| ------- | ----
| Syntonic comma | 220.009
| Ditonic comma | 239.996
| Schisma | 19.987
| Lesser diesis | 420.03
| Greater diesis | 640.04
| Pythagorean limma | 923.002
| Diaschisma | 200.02
| Just fifth | 7180.9997
| Just major third | 3951.989
| Just minor third | 3229.01
| 31-EDO diatonic semitone (3/31 octave)&nbsp;&nbsp; | 1188*
| 31-EDO chromatic semitone (2/31 octave)&nbsp;&nbsp; | 792*
| &nbsp; |
| * Exact because 1023 is divisible by 31. |
<br/>

At <a href="http://xenharmonic.wikispaces.com/"><i>Xenharmonic</i></a>, the same unit has been named the <a href="http://xenharmonic.wikispaces.com/Interval+size+measure"><i>prima</i></a> but not (yet) further discussed. It was Brombaugh's <a href="http://www.tonalsoft.com/enc/t/tu.aspx"><i>temperament-unit</i></a> (tu) that led me to the idea (1&#8239;mil&#8239;&#8776;&#8239;3.00005&#8239;tu).