I agree, the magic-star should generally only identify a sequence that is nicely overspecified. But in the interest of Huffman coding the language I think that if anything warrants an exception, it should be arithmetic sequences (including constant sequences), since these would be the most common use cases (including 1,2,3,4...). Many times I've wanted to iterate over the multiples of N. I think "1,1,..*" or "2,4,6,..*" is just fine, and overspecifying the sequences by one more seems overkill because the rule is so simple/common.
in reply to Re^2: Challenge: Simple algorithm for continuing series of integers (confidence)
in thread Challenge: Simple algorithm for continuing series of integers