Inconsistent annotation of fractions

[mgeorgescu]

The consensus contains inconsistent annotations of fractions:

[1]   consensus   isi_0002.149   (snt. 186 in workset  guidelines, last updated on Tue Dec 11, 2012)
The diameter of Jupiter is about 1/10 the sun's diameter.

(e / equal-01
      :ARG1 (d / diameter
            :poss (p / planet :name (n / name :op1 "Jupiter")))
      :ARG2 (a / about
            :op1 (p2 / product-of :op1 "1/10"
                  :op2 (d2 / diameter
                        :poss (s / sun)))))

[2]   consensus   wb.eng_0003.91   (snt. 91 in workset  wb-eng-0003, last updated on Wed Dec 19, 2012)
How exactly I 'm supposed to keep horses on a 1/4 acre lot is beyond me .

(u / understand-01
  :ARG0 i
  :ARG1 (t / thing
          :manner-of (k / keep-01
                       :ARG0 (i / i)
                       :ARG1 (h / horse)
                       :location (l / lot
                                   :mod (a / area-quantity
                                          :unit (a2 / acre)
                                          :quant 0.25))
                       :ARG2-of (s / suppose-02
                                  :ARG1 i))
          :mod (e / exact))
  :polarity -)

We have also seen inconsistent annotation on consensus and our suggestion would be to stick to the format "x/y" annotated as text.

Typed variants of fractions should also be mapped to the numerical form:

• one third > 1/3
• a fifth > 1/5
• half of the > 1/2

Pls let us know if this looks ok to you.

[uhermjakob]

Good point. I'm not sure whether this was ever documented. I would suggest not to normalize numbers, because some fractions (such as 1/3) do not naturally map to a decimal number and different forms (such as 8 pounds vs. 8.0 pounds) imply different levels of precision.

[mgeorgescu]

We will then stick to using x/y for fractions like x/y and spelled out forms (e.g. 1/3, a/one third). Thank you.