MSC = multiscale clean
MEM  = maxiumum entropy

* A few comments on MEM. I believe that MEM is currently
  over-emphasized in the cookbook -- possibly giving the impression
  that it is the best task to use. For novices MEM is actually
  extremely dangerous -- while you can explain the use of
  the data progress plot for terminating MEM it is more difficult
  to explain the subtlties of masking. Because MEM actually tries
  to force non-masked regions to be smooth -- subtly different masks
  can lead to quite measurably different images -- often with mask
  induced artifacts. Indeed because of this smoothing property
  even the residual map is of limited usefulness if you clean
  very deeply. In contrast the various forms of clean, at the very
  least do not depend that sensitively on the mask -- you cannot
  make real emission go away by not including it in the mask.
  Especially once the properties of the multiscale clean are
  better understood (which removes the annoying clean instability), I
  suggest promoting the various forms of clean and relegating MEM to a
  side note.

* Why are the scales measured in pixels? Wouldn't it be more intuitive 
  in beams? Now if you change your pixel size you also have to change the 
  scales. Using the pixels "convolved with the beam" it isnt' clear 
  to me what PHYSICAL sizes these choices correspond to. If you 
  use a scale (other than delta function) that is smaller than 
  the synthesized beam, what is the resolution really? Doing this 
  is a form of "superuniform weighting" and it is no longer clear 
  what the "real" resolution is. 	

* There are many caveats to choosing these scales that need to be 
  addressed in the cookbook. Conceptually, you should never use scales 
  that are either smaller/larger than your longest/shortest spacing. 
  It useful to have the added warning not to have a scale larger 
  than the mask but for new users, a better discussion of physical 
  scales is needed.

* While the "Total flux" in each major cycle is interesting to see 
  displayed, the number you actually have some idea about is the 
  actual Total flux of the source. It would be nice if the logger reported 
  this as either a running number throughout or at least at the 
  end. Total cleaned flux =""

* If aips++ does not find the mask file, it should complain! I was stuck 
  for an hour trying to figure out what was wrong with my parameters 
  when the problem was actually a typo in the name of my mask 
  file. If you aren't doing interactive, it should fail with a 
  clear message error without a proper mask name.

* Directions for deconvolve single dish for multi-scale clean 
  were hard to follow. Isn't the GBT response incoded psfname='gbt' 
  doesn't work. I looked up psfname in User ref. manual and there is 
  zero information? What is the rational for doing multiscale clean for 
  this? Is it important to pick scales also used for your final 
  combined image?

* When you have an external viewer open, "remove all" does not seem to 
  properly remove the image from memory. i.e. in the case that you 
  display image trial1.im, decide you don't like it, "remove all" in 
  the external viewer, remove file trial1.im from disk and recreate 
  a new trial1.im and display in viewer -- you will be seeing the 
  old image not the new one. A command of "remove all" should aboslutely 
  remove it!