Extended elevation, change of height only

[Olly Betts]

On Mon, May 04, 2026 at 08:08:32AM +0100, Andrew Atkinson wrote:
> On 04/05/2026 04:22, Olly Betts via Survex wrote:
> > owever the orientation of left/right passage dimensions in the extended
> > survey flips if the leg goes into the page vs coming out of it.
> 
> I don't understand. What is the meaning of into and coming out of the page.
> Doesn't left make them go left and right right. Isn't into and coming out of
> the page meaningless. It is relative to something.

I think the confusion here is between folding a leg left/right vs
left/right passage dimensions (LRUD).  Calling this new mode "vertical"
also seems a potential source of confusion - such legs will end up as
a vertical line in the extended elevation, but they're not inherently
"vertical legs".

If you start by looking at a non-extended elevation with the page (or
screen) passing through the start station then the existing extending
process walks the survey network from the start station rotating each
leg in turn so it lies in the plane of the page.  There are two ways to
achieve that, which are called "left" and "right" based on which way
horizontal component of the folded leg heads across the page.

This new "vertical" mode doesn't fold the leg into the plane of the page
though.  It effectively instead rotates it in 3D about a vertical axis
such that it lies in a vertical plane which you're looking at end-on,
then squashes that view to lie in the plane of the page (so the
projected leg only goes up or down the page).

The extending process attempts to also handle LRUD passage dimensions
suitably.  That means they should stay in the same relative orientation
to the legs they are between.  For "vertical" that means the viewer will
end up looking straight on at the LRUD cross-section (at least for a
cross-section between two legs both folded as "vertical" - a cross-section
between a leg folded left and one folded "vertical" should perhaps be at
45 degrees, but let's just consider the simpler case for now).

However there are two possible ways to view a LRUD cross-section which
have U up/D down - the L-R of the cross section could match left and
right on the page, or be the opposite way round.  This effectively
corresponds to whether these "vertical" legs are rotated such that
they head into the page (then LRUD-left is page-left) or come out of the
page (then LRUD-left is page-right).

In a drawn up extended elevation which has been drawn "cut-away" to show
the far wall of the passages, this is like a passage leading off on the
far wall (so has a fully drawn opening) or the cut-away wall (so the
opening is probably drawn as a dotted or dashed outline).

So the question is what to do about this.  The obvious approach to me
is to have (say) "towards" and "away" instead of "vertical", though
people love to moan when Survex doesn't do something the same as Therion
so I'm wondering what Therion does about this.

Perhaps it's possible to pick an orientation automatically based on
angles between the legs at the station where the "vertical" section
starts, though it seems that's going to result in frustrating behaviour
in edge cases (e.g. a slight change in orientation due to loop closure
changes from a newly closed loop elsewhere in a complex cave network
could result suddenly flip which way extend decides to orient the
folding of a "vertical" section) and you'd at least want a way to
override such automatic behaviour.

Cheers,
    Olly