Re: Numbers

[jorge@PHYAST.PITT.EDU]

la kris cusku di'e
 
> >More precisely (but not necessarily more accurately), it expresses a random
> >variable whose range is 2000-2099 and whose measure of central tendency
> >(exactly which measure is unspecified) is 2030.
>
> It would not be a symmetrical function then, with a range and central
> tendency like that.  Jorge's method allows arbitrary specification of either
> the range (20ji'i30 = 20 to 30) or the central tendency (ji'i20 = about 20),
> but not both. Yours allows for both (2ji'i5 = 20 to 29, about 25) but the
> range must coincide with exact powers of ten.
 
Not only it must coincide with exact powers of ten. That wouldn't be such
a big limitation, because at least you could always express the order of
magnitude of the range.
 
The big problem is that you have no freedom at all to choose the range.
It has to be of the "..000-...999" form. This is a completely arbitrary
set of ranges, and if it makes sense to have non-symmetric uncertainties,
then you should be able to choose the non-symmetry. It doesn't make sense
to have only a few non-symmetries available. For a number like "23" you
can only have a nonsymmetry that extends towards the bigger numbers.
For a number like "27", it's the opposite. What's the point?
 
If you ever would need something like "23 with range 20-29", then you
would also need "23 with range 17-26" and "23 with range 22-31" and
all the others. To have only one of them is not much more than to have
none. If such precision is needed, then it has to be available for all
of them.
 
> pe'i for normal inexact human speech John's method is exact enough and more
> powerful than Jorge's, but it's also less intuitive.
 
Could you explain how it is more powerful? That a given expression is
more precise doesn't mean that the method is more powerful. In order
to be powerful you need to be able to choose any expression that has
the same precision. If you only have a limited set of very precise
expressions, then most of the time, you won't have the expresion that
you need.
 
Jorge