http://qs1969.pair.com?node_id=1085828


in reply to Re: At the risk of saying something stupid-but-obvious about Roman Numerals
in thread At the risk of saying something stupid-but-obvious about Roman Numerals

I would have expected to find such a presentation presenting a problem which actually requires a search of an intractably-large space ... and, since this one clearly doesn’t
It does for a golfer. :-) The search is required to improve the previous shortest known Python solution by replacing the shortest known magic formula:
10**(205558%ord(r)%7)%9995
with a new one in this form:
hash(r+"magicstrng")%1001
which is one stroke shorter. This is explained in more detail in the introductory "Lookup Table vs Magic Formula" section of The 10**21 Problem (Part I).

But it would be a lot more interesting if the most-efficient algorithm for solving the actual representative-problem were what is presented “in the least number of keystrokes.”
The shortest algorithms for solving the Roman to Decimal conversion problem have already been presented in: