To mark a rim in decimal fractions takes about 45 minutes, but to do so in spoke intervals takes only a minute. Moreover, readings from spoke intervals are easier to take. The downside is thought to be the time lost in continually in making conversions to decimal fractions, and in the long run more time is lost than with the rim marked in decimals. Pete Riegel reports that most measurers mark with spoke intervals, so they appear to be taking the short-time view.

I now find that with HP RPN calculators such as the 32SII and the 33S that handle fractions, measurers can work with spoke intervals almost as fast as with the decimal equivalents. In fractional display mode (FDISP), spoke intervals are entered directly in tenths and subsequent manipulations yield displays in the same mode. (The decimal equivalents are also readily shown if necessary.)

Take the example of a calibration measurement on a 400-meter course with a 32-spoke wheel that yields 190 revs and the following spoke intervals: 22.4, 23.2, 22.8, and 23.1.

Flags 8 and 9: set.
FDISP with denominator ( /c) fixed at 320.

190 enter, 4 x, 224..320 +, 232..320 +, 228..320 +, 231..320 +, 0.625625 x. Display: 477 84/320 rev/ km. (Constant of 0.625625 factors in the 400 meters, the SCPF, and the 4 measurements.)
5 x. Display: 2386 102/320 rev. For a 5-km course ride 2386 rev and 10.2 spoke intervals.
5 /, 1.609344 x. Display: 768 26/320 rev/mile. Mark miles at intervals of 768 rev and 2.6 spoke intervals.
Original Post

Using spoke intervals is fighting the problem. If one is going to use a counter that counts full wheel revolutions, it makes sense to have the rim marked to give original data in decimal form.

It took me about 1/2 hour to mark my rim the first time. If I had to do it again it would take about 15 minutes.

Ever since I did this I have enjoyed decimal readings, which are far easier to use than 1/32 or 1/36 spoke intervals. Sure, you can overcome the uneven fractions with computer trickery, but why not start with the right tools?

Mark your wheel. You will not be sorry.
Pete - How fine are the gradations you've marked your wheel in? Currently I've got neat little Dymo labels every 0.125 revolutions, and "guesstimate" any reading in between (using the value of the distance between spokes to help).
My wheel is divided into 20 equal parts. They are marked 0, 5, 10, 15.....Getting the first 4 is easy. Those marks all lie at even spokes 90 degrees apart. Then all you have to do is divide the quarter-wheel in between into 5 equal parts. Even if you are off a little bit the error has no bad effect.
I realize now that measurers with ordinary calculators can readily work in spoke intervals with very few extra key strokes. The four spoke intervals from the calibration are added and divided by 128 (144 for a 36spokes) to get the average fraction in decimals. After doing the further calculations in decimals, the final decimal fractions are converted to target spoke intervals by multiplying by 32 (or 36).
I find it hard to guesstimate spoke intervals to less than 0.25. Neville seems to have it down to 0.1 - but is this unnecessary precision? After all, the JO counter only measures 20 or 22 counts/revolution; using half counts doubles that number. Merely using whole spoke counts nearly approaches that; so it would seem half-spokes would be well beyond the precision previously deemed sufficient.
Jim:
I agree that in the measurement of a 5000-m course it is unnecessary to record wheel rotation with the precision that rim calibration allows, because many other factors limit the overall precision of the measurement. However during calibration on a 300-m or less course, it may be useful to do so.
I guess my question is how fine can/should you record spoke intervals or rim fractions? When using the J/O counter, it was full or half counts only. As I said previosuly, half a spoke interval would be 1/64 or 1/72 of a revolution, plenty precise enough, I'd think.