First I want to thank all those who have contributed so far, and may still follow. The collective of responses and thoughts has helped me a lot in getting a better understanding of how people tend to use 'imperial' units of length in various situations.
The biggest revelation was the immense unpopularity of using yards (save some sports related use).
Also the use of feet for very significant distances, but under a mile, is interesting. It seems like people want to avoid decimal fractions, even to avoid using fractional miles. Of course aviation uses feet to indicate even extremely larger altitudes.
The avoidance of using decimal fractions can also be found in using a combination of feet and inches, just use two non-fractional units instead of one. However, people won't panic if fractional feet are used, they just try to avoid it.
Also the context sensitivity can be helpful, e.g. when we want to use these units in Photography related measurements.
I've never heard anyone use fractions of a foot, except for very rough estimates like "5 and a half feet."
Interesting. It's that 'fraction avoidance complex'
I expect the same questions can be asked about the metric system, and answered in a similar way.
Indeed, but I know that better because I grew up with it. In metric, one mainly tries to avoid long sequences of zeros, either before the decimal fraction separator or behind it. It's easier to keep the numbers shorter because they can be read faster and with fewer interpretation errors. It's also why centimetres (1/100th of a metre) are popular, because the size of many objects in everyday life can be expressed in fewer digits when using centimetres.
So, unless your software routine is likely to be used in either an athletic or dress-making situation, I would not bother to include "yards." As for the other measurements, here are two possible approaches:
A. Pick a useful range for each kind of result, perhaps something like this:
If total distance > say 500M, give result in miles and decimal fractions (3.25 miles);
If distance is between, say, 1M and 500M, give it in feet and inches, with decimal fractions of inches (17 feet, 3.25 inches); and
For shorter distances, just give it in inches (27.596 inches).
B. Or, let the user select the range from a menu, including some or all of the following possibilities:
Miles, Miles and feet, Feet, Feet and inches, or Inches.
Eric, that's pretty close to how I had it set up already. I was looking for confirmation on the switching points.
Generally, it boils down to the level and degree of precision you need...
So, I would ask why are you asking? What specific dimensions are you using in what context?
Jeff, I'm writing a tool for photographers (output quality planning), but since the unit converter subroutine (I've pretty much finished it) is generic, it can also be used in other contexts (although I know there are different (international/nautical/survey) miles, depending on context). Feet and inches are officially defined in exact equivalent metric SI quantities, 0.3048 metres and 0.0254 metres (AKA 25.4 millimetres, trying to reduce the number of zeros and digits).
Inches (with decimal fractions) are commonly used to describe e.g. printed output (unless huge billboard sized), or display dimensions. Feet are used for focus distances (also marked on the lens, along with metres), although in photomacrography, inches (and decimal fractions) are more practical (or metric millimetres).
Inches are used to avoid fractional feet. But fractional inches are fine. Inches are handy because thirds and quarters of feet come out as whole numbers of inches.
For dividing miles, it's probably easier to use decimal points (6.25 miles) rather than miles, yards, feet and inches.
David, yes, it seems like the shorter/even numbers are what drives us to a certain degree.
For the rest, tradition dies hard, but for now it seems I'll have to add some sort of conversion capability to my tool to avoid mental disconnect in part of the world ...
I'll probably release my tool with already partial implementation of imperial units soon, watch for the announcement, but full implementation would delay the availability too much, because a significant programming effort and lot of testing is required.
It's not that converting numbers is hard, it's the complexity of the calculations under the hood that must not mix the units up. Some units will remain in metric, like focal lengths in millimetres, and sensel pitch in micrometres, but others may be expressed as various metric or imperial distances. Tricky business, easy to make mistakes, and users can make mistakes as well, so boundary condition checks are part of it all ...