Backsights and default accuracy estimates

[Graham Mullan]

I only have one relevant experience of taking both fore and backsights but
it does throw up some possible problems:

In carrying out the survey we used two sets of instruments (Suuntos) and
leap-frogged the two instrument readers. Thus each set/reader took half the
fore sights and half the back sights. This will mean that unless there is
some way of tagging instruments to readings (and we did not record who took
which) then it is impossible to allocate instrument errors properly. Our
reasoning at the time (this was 1974) was that this technique would
distribute instrument error through the survey. I don't know how valid that
argument was but I do remember being pleased when our results were checked
by radiolocation some time later.

However, given that there may be significant instrument errors (though
having looked back at this data we seem to have had a fairly well-matched
pair [or a lot of luck]) this may mean that the program will trigger lots of
warnings about poorly matched pairs of fore & back sights which are due
solely to instruments with different errors which should have been filtered
out by applying an instrument correction.

I don't suppose the above will help solve any problem but it might help
point up some of the additional complications that arise from adding
additional readings.

Graham
--
graham.mullan@breathemail.net

Tel 0117 9502556 or 07887 637064

Information in this message would be confidential if I were not sending it
out over the Internet! If you are not the intended recipient, I hope you do
not find it overly boring.