[evla-sw-discuss] Reference pointing

[Barry Clark]

> is there a clear advantage to bundling them over always keeping them 
> separate?

It is a little cleaner, I think.  You don't have to use the fancier stuff
unless you really want to.

Instead of
z = subarray.refPointingInit()
subarray.setRefPointing(z)
<X band pointing scan>
refpointingAgent.register(z)
<short dummy scan to wait for results>
<pointing scan at K band>
w = subarray.refPointingInit(z)    # copies X band offsets
refPointingAgent.register(z)
<K band target obsn>
subarry.setRefPointing(w)
<pointing scan at Q band>
refpointingAgent.register(w)
<Q band target obsn>
etc.

you can use the simpler

z = subarray.refPointingInit()
subarray.setRefPointing(z)
<X band pointing scan>
refpointingAgent.register(z)
<short dummy scan to wait for results>
<pointing scan at K band>
refPointingAgent.register(z)
<K band target obsn>
<pointing scan at Q band>
refpointingAgent.register(w)
<Q band target obsn>
etc.

if you are content to start the Q band pointing peakup from the K band 
position instead of the X band position.

> i also don't quite get why the dummy scans are needed?

The offsets are not available until the pointing scan is completed and 
reduced.  If we start a new pointing scan immediately, the pointing offsets
will arrive after we have set up for the first offset.  This might seriously
confuse the pointing solver.