Should decimal be the default for floating period literals?

E.g.
1.2 would actually be decimal.Decimal("1.2")
and float(1.2) would be used to get traditional binary float point.