Autopilot at Zero Longitude.

[AntarcticPilot]

I can't understand a software problem at the 0⁰ longitude - everything behaves as expected there; divide by zero or NaN shouldn't arise. Problems arise at 180° where you go from +179.99999° to -179.99999°, though, and older Raymarine kit has a known problem there (it selects features for display using a box that goes all round the world, not across the 180° meridian!)

 

[Lucky Duck]

A few boats back, the GPS repeater cockpit display would show 000 degrees and 360 degrees

 

[dunedin]

I can't understand a software problem at the 0⁰ longitude - everything behaves as expected there; divide by zero or NaN shouldn't arise. Problems arise at 180° where you go from +179.99999° to -179.99999°, though, and older Raymarine kit has a known problem there (it selects features for display using a box that goes all round the world, not across the 180° meridian!)

Yes, and because the Greenwich Meridian is on a very busy yacht and ship route crossed very frequently, any generic software bug would have been uncovered almost immediately. Sounds like the OP had a different issue or coincidence.

 

[mjcoon]

I can't understand a software problem at the 0⁰ longitude - everything behaves as expected there; divide by zero or NaN shouldn't arise. Problems arise at 180° where you go from +179.99999° to -179.99999°, though, and older Raymarine kit has a known problem there (it selects features for display using a box that goes all round the world, not across the 180° meridian!)

That's a very simplistic viewpoint. We don't know what the plotter was trying to calculate, bearing to a waypoint, or whatever. Presumably these days they work solely in WGS84 so do not have to do the challenging 3D geometry calculations to get to other datums...

 

[AntarcticPilot]

That's a very simplistic viewpoint. We don't know what the plotter was trying to calculate, bearing to a waypoint, or whatever. Presumably these days they work solely in WGS84 so do not have to do the challenging 3D geometry calculations to get to other datums...

Simplistic works in this case. I implemented this kind of software for many years, and while I encountered problems aplenty at 180°, I never had a problem at 0°. Problems at +90° and -90° latitude also happened. But at 0, all the relevant functions are continuous and well behaved.

I agree that datum shifts are unusual these days, but many GPS systems still have the ability to use datums other than WGS84; OSGB36 and ED50 come to mind. It's something I've never had to implement, but I think it's a straightforward set of equations using a set of up to 11 coefficients.

 

[justanothersailboat]

I used to implement geographical software on a much, much humbler level than AntarcticPilot (as an approximate analogy to the level of seriousness, think him:"your charts" vs me:"your school atlas" ) and honestly I'd be unsurprised, but really disappointed if commercial navigation software had any trouble with any of this. It's basic edge case handling, so what if the angle is 361, the software should take it in its stride.

(from another job in another field, I learned that "unit test" is a quite rude word in the language of a country where a lot of software is outsourced to, so who knows)

 

[Frank Holden]

The seals may not be fat for much longer. Those waters have been overfished for decades, it will be as dead as the Atacama before too long.
You can see the mainly chinese fishing boats in the two attached screenshots, second one is zoomed in to the waters off the Peru/Ecuador border to show the actual numbers. Fishing the living daylights out of it just outside the 200 mile limit.

And don't get me started on the rape of the South Pacific mackerel fishery. It takes 4kg of fishmeal made from mackerel to produce 1kg of farmed salmon.