[Gate-users] Ring repeater - request for clarification

[Paul Betteridge]

Hi,

Anders Garpebring sent a query to this list last week about the ring 
repeater, and I have spent a little time investigating why it seemed to 
work for me and not for him. The behaviour I found is unexpected (at 
least to me), and I'd appreciate if someone could comment on it.

The rotation axis for the ring repeater is specified by the setPoint1 
and setPoint2 commands. I have tried to use it to specify both the 
direction and the position of the axis, like this:

/gate/Name_Volume/ring/setPoint1 0. 1. 20. mm
/gate/Name_Volume/ring/setPoint2 0. 0. 20. mm

which corresponds to the User Guide:
http://wiki.opengatecollaboration.org/index.php/Users_Guide_V7.1:Defining_a_geometry#Repeating_a_volume

This seems to work, and it behaves as I expect. But if you change the 
unit from "mm" to "m", it doesn't work any more.

Now in fact, the behaviour of the setPoint1 and setPoint2 commmands 
defined in the GateAngularRepeaterMessenger class doesn't specify a unit 
to be read with the 3 numbers. This is the code:
https://github.com/OpenGATE/Gate/blob/develop/source/geometry/src/GateAngularRepeaterMessenger.cc#L35

which can be compared with the code for the linear repeater, which does 
have a unit:
https://github.com/OpenGATE/Gate/blob/develop/source/geometry/src/GateLinearRepeaterMessenger.cc#L29

So, the actual behaviour of the ring repeater seems to be that you can 
use setPoint1 and setPoint2 to specify the direction _and position_ of 
the rotation axis in _millimetres_.

I have also noticed that GateAngularRepeater::DescribeMyself() seems to 
assume distances in cm. See:
https://github.com/OpenGATE/Gate/blob/develop/source/geometry/src/GateAngularRepeater.cc#L131

But that code doesn't actually seem to be connected up to the /describe 
option for the ring repeater - I haven't tracked down what is wrong.

So my questions are:
  * is the behaviour that setPoint1 and setPoint2 don't read in a unit 
deliberate or accidental?

  * is the way that both Anders and myself have tried to define rotation 
about a point acceptable?

(I would imagine so, since the calculation is delegated to 
HepGeom::Rotate3D()).

  * if the existing behaviour of the repeater is accidental, is anyone 
interested in trying to fix it?

I'd hesitate to fix the code myself, because I suspect there may be 
other consequences I can't see, but I would be happy to make the 
documentation reflect reality.

Thanks,

Paul

-- 
Paul Betteridge
Adaptix Imaging

paul.betteridge at adaptiximaging.com