Here is what happens with a real project that I did a year ago.

As far as extended elevations go, it is a well-behaved project, that was built incrementally, starting with survey 1, then 2 etc.

While the extended network behaves well, the log does not do so well.  Here is a small sample.

Curiously the extend log shows it starting in the middle of the survey sequence (it goes to survey 6 or 7) at a random entrance.  There is no start 4.24 in the data, only a start 1.0!

And yet…

Two more things:

The sequence from 4.20 to 4.19 all the way to 4.15 is extending right, but the log says it is left.

I deduce that the sequence then steps back along the same network until it is time to extend along a new branch.  That is what the BACK refers to (nice).

Here is what the specified extended network that generates both of the images above look like.

and much further down the table it gets to the bit of passage in the image above…

So, the extend log, although following a credible path along the survey network, has some directional errors, and is nothing like the specified path.

Curiously, this log, like the logs for some other projects of mine, end with some apparently random non-existent STARTs and stations that don’t necessarily exist, like -@1 in the sequence below.

So, this feature has real promise, but I wonder if there is a bug, and now I am really interested how the extend specifications and logs compare in Tarquin’s short oxbow examples.

Bruce