[Hist] Fuzziness

Had a think about this, and I think the most natural way is just to provide "error bars" in the same format as the date, in the same units as the most detailed part of the date. A blank fuzziness counts as 0.

So:

-794 with fuzziness 5 means 794BC (but could be 799-789)
-794 with fuzziness 0 means 794BC (some point in that year)
55-04 with fuzziness 2 means Feb-Jun in AD55 (i.e., spring 55)

Note that this last one is actually less fuzzy than the previous, despite having a non-zero figure.

There are (occasional) cases where you want different fuzzinesses at opposite ends of a date range, which I suggest we do as something like 5/1. (The obvious example is Saul's reign, where the end is relatively clear but when he started is a bit of a guess.)

There may be cases where the fuzziness differs between the early and late Exodus dating schemes (in the Judges, for instance), but we can decide what to do about that as and when we find one.

Make sense to everyone?

Colin