Mangled 3d file

[Wookey]

+++ Olly Betts [2015-03-04 02:16 +0000]:
> On Tue, Mar 03, 2015 at 04:53:44PM -0600, Bill Gee wrote:
> 
> The default isn't to require an exact match - the tolerance of the check
> is based on the SDs set for the fore and back instruments, and the
> default SDs are 0.5 degrees, so the test is based on a combined SD of
> 0.707 degrees (based on the assumption that the fore- and backsight
> errors are independent).
> 
> The check currently seems to be if we're outside of 2 SDs, which if the
> SDs are correct you'd expect to be true for 5% of non-blundered
> readings:
> 
> http://en.wikipedia.org/wiki/68%E2%80%9395%E2%80%9399.7_rule
> 
> So that translates to a default threshold of 1.414 degrees for the
> difference.  Unless you read the instruments to better than 0.5 degrees,
> that means the threshold is effectively 1 degree.
> 
> > I don't want to see 200 warnings for readings that are within the
> > tolerance we chose.
> 
> Wookey also commented a few days ago that these warnings were very
> noisy.  I've not dealt with much data involving backsights, so it would
> be useful to get feedback if these warnings are too readily triggered.

The tolerance used on Erin's China trips is 2 degrees, so most readings are within 2 degrees.

As you show, this doesn't exactly match the *SD-related warning,
although it's similar.  I wonder if one should change the SD to fit
the 'expected' tolerance, or add an explict 'tolerance' warning
threshold? The two numbers are related but not actually identical.

A large project inevitably collects a lot of fore/back sights that are
out of tolerance over time, and thus compiling a big system gives a
lot of warnings, which obscures genuine errors in the new survey you
just added. Perhaps just having an option (like gcc) to turn off
specific warnings is one way to deal with this?

> On the 107 legs in AllieSpringCave.svx, the default SDs give 35
> warnings, though some of those look like genuine issues (e.g. a 7 degree
> difference).  It seems to be noisier than expected, probably partly due
> to the effective threshold being lower.  Perhaps also 0.5 isn't entirely
> realistic for the SD of a sighting compass reading.
> 
> I also suspect a 2 SD check probably isn't the right criterion anyway,
> as even 5% false positives is unhelpful.  A 3 SD check would be expected
> to give a false positive about once every 370 readings (or 185 legs),
> and for the default SDs would mean a threshold of 2.121 (effectively 2)
> degrees.

That would fit the hmg dataset nicely. 
 
> When surveying with sighting instruments using backsights, what
> threshold is generally regarded as acceptable?

Ask on an American survey list. But I'm familiar with 2 degrees, Bill
says 1 degree on his project. 

For actual SDs I'd say that (well-calibrated) distoXs have better SD
than sighting compasses, especially at high angles, but if you end up
with a pair that are not well-calibrated it can be a struggle getting
to 2 degrees. (Sighting compasses should strictly have an SD function
that takes into account the clino angle (as opposed to a constant)- so
there is a step at 15 degrees where it starts to stick, then it rises
to several degrees at +89 clino).

Wookey
-- 
Principal hats:  Linaro, Debian, Wookware, ARM
http://wookware.org/