# Area 51 proposal: Mathematics golf

I recently (i.e. yesterday) put a proposal for a 'Mathematics golf' stack exchange on area 51. This would follow the same concept as 'code golf' for which there is already a stack exchange for: https://codegolf.stackexchange.com/. Please let me know your thoughts / support it if you think it is any good.

EDIT-in response to Jyrki Lahtonen's comment

Here is a quick summary of the concept of code golf:

1. A problem/challenge is set in the question, to be solved using a computer program (e.g.: find all the prime numbers under 10,000).
2. The answers given contain peoples attempts to solve this problem.
3. The answers are scored based on the number of bytes they contain.
4. The answer with the lowest number of bytes 'wins'.

The reason for the word 'golf' is because apparently golf has a similar scoring system.

The proposal given for a 'Mathematics golf' would be along the same lines, i.e. a problem (e.g. a calculation or derivation) is set to be solved using mathematics with answers containing peoples attempts. A similar scoring system could be used as in 'code golf' so that the shortest wins.

• This could be interesting. Is the idea that the shortest way of proving/calculating whatever is asked is the "winner"? IOW, can you explain the concept a bit more verbosely to those who have never visited codegolf, or only taken an occasional peek. – Jyrki Lahtonen Apr 5 '16 at 8:48
• @JyrkiLahtonen I have edited a response into the 'question'. – Quantum spaghettification Apr 5 '16 at 9:03
• Can you objectively measure shortest solution on mathematics? For example two solutions of this form are basically the same: "A=B, which obviously implies C=D" and "A=B, and therefore (insert a proof long several lines here) C=D" are basically the same. I could imagine using some software like Coq or other theorem prover to measure length of proofs, but I suspect not many people would want to write proofs in this way. (Certainly I would not be able to write formalized proofs in some of these programs.) – Martin Sleziak Apr 5 '16 at 10:04
• When I saw the post I thought "great idea" but then I read the details and now I fear this is too narrow. For one thing, tt seems like a subset of codegolf and it is not quite clear there is a need for a dedicated site on this. A site for mathematical puzzles and contest, including but not restricted those you mentioned, would seem more viable to me. – quid Apr 5 '16 at 10:32
• @quid I agree, with our first comment, that a 'site for mathematical puzzles and contest' would be better. Although this may then be similar to 'puzzling' stack exchange. – Quantum spaghettification Apr 5 '16 at 10:46
• @quid Maybe I misunderstood the post, but the part mentioning computer programs is description of codegolf. About this proposal the OP says: "problem is set to be solved using mathematics with answers containing peoples attempts". Also looking at example question, I do not think that answers are supposed to be computer programs. – Martin Sleziak Apr 5 '16 at 10:53
• @quid So, if I get it correctly, there would be mathematical questions that can be answered using a computer program and the shortest program wins. But it seems that there is already a (math) tag on code-golf. – Surb Apr 5 '16 at 10:54
• @MartinSleziak no, it was I that was careless in reading the post. Sorry! I removed the comment. – quid Apr 5 '16 at 10:58
• @Surb this is what I thought, but maybe the proposal is broader. Anyway I think it would be better to have it still broader. – quid Apr 5 '16 at 11:02
• @Quantumspaghettification AFAIK puzzling does not welcome math problems only math puzzles. And the distinction is such that I think there is some room between Puzzling and Mathematics that could be filled. – quid Apr 5 '16 at 11:05
• @quid just to clarify, I don't mean that the answers should be computer programs, they can be normal mathematical expressions. (I was just giving a description of code golf when I mentioned computer programs). – Quantum spaghettification Apr 5 '16 at 11:12
• @Quantumspaghettification Ok, but (as already pointed by Martin Selziak) how do you measure the length of a proof? Maybe, you could explain what kind of answer you would expect to your sample question in the proposal and what is the objective. – Surb Apr 5 '16 at 11:14
• It seems to me that this proposal could serve purposes similar to what was previously discussed here: Using Math.SE for contests, Competitions on MSE, Are small competitions allowed? – Martin Sleziak Apr 5 '16 at 14:30
• Maybe the scoring could be based on the number of calls to a given set of axioms for solving the question. – Surb Apr 5 '16 at 14:38
• @NajibIdrissi If you have a look at some theorem proved using Mizar or at the list of theorems here (link taken from this paper), it seems that there are some non-trivial results with completely formalized proofs. Of course, they were checked using computers on by hand. Here on math.SE similar problems seem to be usually in first-order logic. – Martin Sleziak Apr 7 '16 at 4:54