badger
2007-10-11 10:19:50 |
solution methods/labels
I am new to this site. I see that some postings of Solutions are marked, e.g. (for most recent puzzle) "full solution (explanation and answer)" or "computer solution (spoiler)".
From other sites to which I post I know there are sometimes extensive discussions of the relative merits of solutions which are derived by logic/inference alone, and those which use computer tools as well (assuming that the solutions by both are correct). Does this site AS A SITE make such a distinction in preferences or ratings (as distinct from the preferences of individuals who may post here) ?
The mixtures between inference and algorithms may vary with the problem, and within a problem between individuals. In any case, I find the most valuable and interesting are those which give explanations, and not merely answers. Of course a certain amount of inference is necessary to understand the task at hand to create an effective and efficient program.
This knaves puzzle is a good case in point. I wrote a short program to test all 27 assignments of three individuals to three types, and for each of those to score the 8 statements; with those results at hand I used non-computer reasoning to come to the solution (though that step also could have been programmed).
What views (or rules?) --- ?
|
brianjn
2007-10-11 20:41:16 |
Re: solution methods/labels
You raise several issues here.
When comments in the solution area are posted some people will choose to suggest that they have the solution by means of the light bulb. This can be quite misleading as sometimes some such comments are not solutions at all, and in some instances some of us will post a solution and not declare it as such so as to 'goad' others into seeking their own solution (without maybe looking at ours); gamesmanship, one might say.
Solution techniques can be a bane. It is rather frustrating when you have a puzzle which deserves (in your opinion) the beauty of reason and a computer generated solution emerges.
In such circumstances it helps to include some form of wording as to what you as the author require, eg, "how would you derive...". Sometimes a comment like, "without calculating devices ...", is used.
One of our solvers is rather adept at providing computer solutions, and quite decently provides explanations as to the general thrust of the program and its interpretation. The essential part of the program listing is also provided. A few others have offered similarly.
As to the arrival of logic/inference solutions you will note that regularly someone posts a bland stated solution without any derivation or supporting comment; often these are challenged.
Really there are no hard and fast rules. Generally an author's requests will be treated with respect. I once had a problem where a computer generated solution was highly likely, that did emerge, but after a time when the puzzle had been exposed to several commentators.
Hope that helps. |
