Hi,

The results formatting of engineering computations should reflect their valid precision. Typically, this is assumed to be tree significant digits, perhaps 4 if the leading digit is 1. This is based on the precision of most available input data and in former times that was the achievable precision of the slide rule.

I find myself adjusting the number of digits in the right button menu on a case by case basis and have to correct the settings if Input data changes.

Currently, the only way to enforce a given relative precision independent of the result's order of magnitude is to set the exponential threshold to 1 and to specify the number of digits accordingly. However, the exponential threshold is better used to provide readablity than to control precision.

Very often I get results displayed as zero just because they are smaller than 0.01 and default precision set to 2 digits.

One can use brute force to set the relative precision by defining a function like

format(x_,n_):line(oom:round(log10(x_)-0.5,0),round(x_/{10^oom},n_-1)*10^oom,2,1)

This yields

format(1234563,2)=1200000

format(0.00001234,2)=0.000012


Proposal:

• Provide a control over the number of significant digits including cases like 1234556 -> 1230000.
  
• Include an option to increase it by one if the first digit is 1. Setting "3" would then provide 1234000
  
• Include a setting of the global default for the option to retain trailing zeros.
  
• Provide control over corresponding default settings

My hope is that this might be useful for other users as well and that it is not too complex to implement.

Best regards, Martin Kraska

Edited by user 27 March 2014 16:07:41(UTC)  | Reason: Found out how to swith on display of trailing zeros in the context menu