Survex 0.99 prerelease 5

[Olly Betts]

On Tue, Nov 06, 2001 at 08:28:02PM +0100, Andy Waddington on Survey stuff wrote:
> I use .pos files when I want to know the locations of just a few stations
> from a large survey. This happens on several occasions:

If you just want to find coordinates for a few station, Aven provides two
easy ways to do so - you can point at a station on the survey, or find the
station in a hierarchical survey tree.  Probably easier than rummaging
through a text file for them.

Alternatively, if you're wanting to read station positions automatically
from a program, and don't want to use the img library, then just get your
program to run 3dtopos on the input .3d file, and then read the pos file it
produces.

> 4) When "inventing" survey data for bits of cave which haven't been
> surveyed, but which have some details available (like pitch lengths).
> Here you need to know the locations of the two ends in order to invent
> "joke" survey data which doesn't distort the rest of the cave when added
> to the dataset, but which *does* appear to connect (eg. the "fiction
> survey" in 1623/41).

Never, never, never do this.

You might initially get joke survey data which doesn't distort the current
dataset, but any additional data added to the dataset which changes the loop
closures will mean the joke data no longer fits exactly, and this misfit
will be distributed back across the rest of the data.

The best approach is to make a leg in the joke survey a "nosurvey" leg -
such a leg has a FROM station and a TO station and no other data - the leg
automatically fits the gap between the two stations.

OTOH if you do have a survey but it is less accurate than the other data,
then all you need to do is set the accuracy and the network solver will
distribute the errors appropriately.

> None of these usages is necessarily "mainstream", but all are much more
> easily done with the sort of output you get from a .pos file. For some of
> these uses, you probably don't know in advance that you will be wanting a
> .pos file, and it is just simpler to have a .pos to hand when the need
> arises.

Um, to me this suggests not adding a switch to produce a .pos file, which is
what I thought you were arguing for.

> I can envisage situations when you no longer have the option of
> rerunning survex on the same data (a good reason for keeping the data in
> CVS, so you can go back :). Perhaps a separate tool to generate a .pos from
> a .3d would be handy if .pos files are not to be generated by default.

You mean a tool like 3dtopos?

Cheers,
Olly