Grid convergence

[Olly Betts]

On Tue, Feb 16, 2016 at 02:20:47PM +0100, Mateusz Golicz wrote:
> Let us consider this test. Suppose an unsuspecting user surveyed a
> very straight Chinese borehole in Austria. It starts with a huge
> entrance at station "0" and ends in a choke at station "1".
> [...]
> The way UTM works is that if you go 1000 m towards the
> north pole from 328000 5260000, you in fact are roughly at 328029
> 5260999, you like it or not. This is some 29 m away from what Survex
> indicates.

Shouldn't be a problem - a proper Chinese borehole passage would be
at least 58m in diameter.

> Of course this "entrance hunting" example is a bit strange, but I
> guess we all realize this has big consequences in closing loops in
> multiple-entrance, multiple-fix situations. How do "we" deal with
> this? Any plans to fix that? If not, shouldn't we at least warn the
> unsuspecting user in our manual?

It's something I'm aware really ought to be handled, but I haven't
looked much at how best to do so.  I know therion has code to deal with
grid convergence, so doing something similar might make sense.

For the magnetic declination, the main difference in the approach I
took for Survex is that you have to explicitly specify a location
for the declination calculation - if I follow therion's code, it uses
the average of the fixed points as an implicit location, but that means
the declination is affected by adding a fixed point, which seems
surprising and unhelpful.

Survex's approach also allows specifying declinations for different
parts of the survey, which is good if you are putting in data for caves
over a wide enough area that the declination varies significantly.

It seems therion calculates convergence at the same point, so I think
it probably makes sense to handle this in the same way as the
declination for Survex.

Cheers,
    Olly