[daip] Re: aips max # stations

[Eric Greisen]

Chris Phillips writes:
 > Hi Eric,
 > 
 > I am interested in using aips (UVCON) for some SKA simulations. If I
 > understand correctly aips has a maximum # stations of 255 due to the way
 > station numbers and encoded in baselines.
 > 
 > Is this correct? I need >300 stations. I also assume this is a pretty
 > fundamental limit (no recompiling with a changed .inc file to increase the
 > number).
 > 

This is a very deep problem - I think UVCON ignores it at some level
but then baseline selection etc does not work.  I think SKA
simulations have been done but we have a problem with th3 whole
concept of including a baseline number as n*I + J - one can easily
exceed the accuracy of counting with floating point.  Sault introduced
a concept of using a dual system with a larger n - if the "baseline
number" > his larger n then it is his system else it is the old system
- but I hate to tell you how many 1000's of line that would affect.
And his schem still does not cover the needs of some SKA models.

Eric