[alma-config] Difference images for the simulations using theSIL

[John Conway]

> 
>     I feel the approach of using the full resolution image to compute the UV
> data set is closer to reality.
> However, it should not matter so much which approach is used if sampling is
> adequately preserved.
> 
>         Hope this can be useful for you
> 
>         Regards,

I see your point about being closer to continuous 
reality, but even in the downsample first before generating uv approach
there are always 4-5 pixels per beam so as you say the sampling effects
should be small, and at least no larger than in the A-array
simulations.

Since this registration problem has held up the AIPS sims for a month 
a two already, - and we don't want to wait even longer maybe Steven should
go ahead and generate the B,C,D results quickly with the downsample first
approach and then while  we are looking at the results try to apply the
strategy you have develeped for your GILDAS sims to his AIPS ones. 

  John.