4 ms·
Interesting to note that decimal fractions (e.g. 0.3) are awkward to represent with floats in computers, whereas fractions based on 1/2, 1/4, 1/8, etc. that are
by billti 6y ago
Interesting to note that decimal fractions (e.g. 0.3) are awkward to represent with floats in computers, whereas fractions based on 1/2, 1/4, 1/8, etc. that are often used in the non-metric system are a very natural fit for the float representation.
- gspr 6y agoNothing stops you for using fractions with metric.
- war1025 6y agoNothing stops you from using decimal in Imperial :)
- gspr 6y agoTrue, but: If you use decimal with metric, the units and prefixes play along with you very well (huge advantage). If you use other fractions with metric, you get no special bonus or burden (neutral). If you use decimal with US customary ("imperial"), you get no special bonus or burden (netural). If you use other fractions with US customary, you get no special bonus or burden (neutral). In summary, there is absolutely no rational reason to stick with US customary, besides legacy compatibility (technical debt).
- JdeBP 6y agoBut, conversely, converting 3 pounds to stones with floating point is not quite as natural. (-:
- brianwawok 6y agoBut literally no one uses stones. People in UK when measuring weight of humans? I have never seen it in any other case. You don't buy things in the US by Stone.
- JdeBP 6y agobillti didn't say anything about who uses things, but you can enjoy how similarly unnatural binary floating point is for converting 8 inches into feet if you like, or 2 fluid ounces into U.K. pints.
- ardacinar 6y agoThat would be a good point if the non-metric system we are discussing was based in binary. It's mostly (but not entirely) based around base-12 and uses 1/3 very commonly as a fraction (as well as 1/6). 1/3 and 1/6 does not have an exact floating point representation.