Yes, the reason I answered your post was that I thought other monks might make the same mistake I initially made: that's why I thought it was worth showing that your solution worked for the sample data shown because of numeric coercion, and also to show the test.

This being said, and as pointed out by AnomalousMonk in Re: sorting an array with decimal points, if the OP's data is really representing version numbers (and I think that's what they are) and not decimals, then your solution would fail on, for example, 5.2, 5.14, 5.16, 5.20, because it would sort 5.2 after 5.14 and 5.16 (which would be wrong for version numbers).

But, of course, my assumptions that they are version numbers may also be wrong. Only levW can tell us what they really are.