NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Posting Statistics / Proposal
From: Chuck Taylor
Date: 2001 Jul 24, 11:33 EDT
From: Chuck Taylor
Date: 2001 Jul 24, 11:33 EDT
Another poster pointed out that simply averaging the latitudes and longitudes may not result in the most centrally located place for the meeting proposed by Dan Allen. Here is another proposal for selecting a meeting location. 1. Given the lat/lon of two points, it is relatively easy to compute the great circle distance between those two points using the law of cosines. 2. Minimizing the average great circle distance between the meeting place and the the homes of list members is a reasonable criteria for selecting a meeting place. 3. We are not likely to meet in the middle of the ocean, but rather at one of a finite number of coastal or island locations such as Salem (Massachusetts, USA), New York, Miami, Greenwich or London, Lisbon, Copenhagen, Cadiz, Rome, Athens, the Azores, Bermuda, Rio de Janeiro, Capetown, Sydney, Auckland, Hong Kong, Honolulu, San Francisco, or Seattle. (I'm sure readers will nominate other sites.) 4. Here is the algorithm, which would be fairly easy to program on a computer (assuming that you remember to convert between degrees and radians :-) ): a. Let A1, A2, ..., AN be the locations (lat/lon pairs) of list members. b. Let B1, B2, ..., BM be the locations of possible meeting sites. c. For each i in [1:M], compute the great circle distance between site B[i] and each of the A[j] for j in [1:N]; average those distances. d. This will produce a single number (average great circle distance) for each of the candidate locations. Pick the site with the lowest average great circle distance. Any takers? Chuck Taylor 47d 55.16N 122d 11.18W