NavList:

Antoine, you wrote:
"Well, I think that rather than an "operationally oriented" one, Frank's last post is essentially a well detailed "didactic/educational" post, and I appreciate its value on this specific ground."

I should also add that there are indeed people who have done just what I described. It really only applies to people who clear their sights with software, but there are plenty of those these days. Luis Soltero who created the very versatile "Starpilot" software for the TI-89 calculator once commented that it was "surprising" to him "how rapidly the sights converge on the GPS position" (as a calculator fan, Antoine, you would probably enjoy his software though it's a bit pricey if you have no practical need for it). For a scenario, imagine you're in the Caribbean, and you want to shoot some Sun sights. Maybe you fried your main GPS system, dropped the spare GPS batteries overboard, but the nav calculator is still running, and you can't bear calling for help quite yet. So you shoot the Sun and enter the sight data. You tend to other tasks, spend a few minutes digging around hoping maybe there are spare batteries hidden away somewhere, then you shoot the Sun again five or ten minutes later and enter the sight data. You keep doing this for an hour or more. It occupies only a small fraction of your time. The question here is how long does it take for this sort of "rapid-fire fix" to converge on your position in a "useful" way, whatever useful might mean. It's a problem that is relatively difficult to work by plotting LOPs, since they overlap so, and because of this, many navigators and navigation enthusiasts assume it can't be useful.

You wrote:
"And no doubt in my mind whatsoever, that if he certainly would like to spend most of his days holding his beloved sextant when at sea, Frank does give due priority to more immediate tasks."

Just so there's no misunderstanding, I get to sea "as often as a Chauvenet", to coin a phrase. But I thank you for the sentiment!

And you wrote:
"Frank, you certainly must be a GOOD instructor, because I really and definitely enjoyed reading your last post on the very clever way(s) to reduce the size of an ellipse!"

Thanks! Glad you liked it. George added a mildly mocking comment about the issue of the 95% level chosen for the size of the ellipse. Just so we're clear, OBVIOUSLY choosing a different probability level does not improve the fix. The error ellipse is what it is. The point is that this particular probability level, 95%, is a purely arbitrary choice. One could just as easily pick some other level so long as we make the same choice for all error ellipses. If you find that an error ellipse has dimensions of 1 n.m. by 10 n.m. this doesn't mean much by itself. We need a common standard error ellipse to compare against. I suppose the best choice for a standard would be the error ellipse for a two-body fix with the objects 90 degrees apart in azimuth for some set standard observational error since that's the best minimum fix.

You concluded:
"I now know the "n^3/2" method."

To avoid any future misunderstanding, this variation with n applies to that specific example. There's a longer formula for the more general case.

-FER

----------------------------------------------------------------
NavList message boards and member settings: www.fer3.com/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]fer3.com
----------------------------------------------------------------