[Therion] therion.log: What does "Survey has 2 connected components" mean?

[Olly Betts]

On Tue, Feb 09, 2010 at 10:10:24PM +0100, Torstein Finnesand wrote:
> Below is a part of the therion.log file. It says "Survey has 2 connected  
> components."  What is this?

It means (Survex thinks) there are two survey networks which are connected
within each other, but not to each other.  So two separate caves, each
fixed by a GPS fix for example.

But your survey has no fixed points (just an invented one) so that shouldn't
work.  You need at least as many fixed points as connected components.

> It also says that there are 87 loops. But when I count them in  
> therion.log, there are 86 (which should be correct, I think: 
> 3007-2922+1).

The formula used is:

    #legs - #stations + #connected_components

So the bug here is that it is deciding there's a second component.

> If I rearrange (sort) the centerline data, and compile again, - the  
> sentence is gone. It says that there are 86 loops (a part of that  
> therion.log is shown below s well).

There must be something about the first arrangement which the colouring
algorithm used to find the components isn't handling.

If you're willing/able to send me the first dataset, or a cut down version
which Survex still incorrectly claims to have 2 connected components, that
would be very helpfully in tracking this down.

Cheers,
    Olly