Motion M0046.02: Literals clarified: YES
I vote YES on Motion 46.
I wonder whether it would be useful to mention that the "Uncertain form"
is adequate for narrow (and hence bounded) intervals not containing zero,
but is severely restricted otherwise.
I suppose the integer ulp-count can have more than one digit; the examples
all have a single digit. If ulp-count has at least as many digits than the
midpoint (or endpoint, for asymmetrical forms) it is possible to represent
intervals containing zero, but that is an awkward use of the notation.
Note that the ulp-count cannot itself have an exponent; it inherits the
exponent from the mid-or-endpoint.
I note that if we had a notation for an infinite radius, the up or down
forms could represent semi-bounded intervals. In fact, if a missing "r"
were interpreted as oo instead of 1/2, the current syntax would suffice!
I also note that the Vienna proposal allowed a non-integral ulp-count,
but its examples were also just single-digit ulp-counts. This was less
limiting because Vienna also had explicit midrad support (<m+-r> with m
and r both floating-point literals), as well as relative and absolute
uncertainties (where the part that took the place of ulp-count was shown
as a floating-point lietral without exponent part).
Michel.
---Sent: 2013-07-29 21:01:24 UTC