Sergey Karavashkin: 
 >dubious@radioactivex.lebesque-al.net (Bilge) wrote in message:
 >> Sergey Karavashkin: 
 >>  >dubious@radioactivex.lebesque-al.net (Bilge) wrote in message
 >>  
 >>  >>   Just post the function. If it exists and you know what it is, then
 >>  >> nothing prevents you from posting it, especially given the fact that
 >>  >> you would be quite happy to rub my nose in being wrong for telling
 >>  >> you that \nabla x (\nabla\Phi) = 0 is a mathematical identity.
 >>  >
 >>  >;-) 
 >>  >
 >>  >http://angelfire.lycos.com/la3/selftrans/v4_1/contents4.html#dipole
 >>  
 >>   Just post the function. I'm not going to read nine pages of a convoluted
 >> argument about the curl of a gradient not being zero when you can post the
 >> function for which you claim that relation doesn't hold in a single line.
 >
 >Remain ignorant. Your difficulty. ;-)

  You misspeled ``logical'' in that first sentence. It's a simple
question to resolve, sergey. 

  (1) Either you have a function f for which curl (grad f) is non-zero or
      you don't.
  
  (2) If you have such a function, then either the function is nine pages
      long, or I don't have to read a nine page argument to take the curl
      of the gradient of that function.

  (3) If I take the curl of the gradient of that function, I will
      either get zero or I won't.  

  (4) If I get zero, then you're wrong in claiming the curl of a
      gradient can be non-zero.

  That's all there is to it. It's not rocket science. Add to this,

  (5) If you had such a function, you would waste no time in posting
      it, if for no other reason, to force me to admit you were
          right and that curl (\grad f) = 0 is not a vector identity.

  I conclude that despite a lot of motivation to post such a
function, you haven't done so because you know that if I take
the curl of the gradient of that function, I'll get zero without
inserting your nine page argument somewhere between taking the
gradient and taking the curl which magically alters the vector
function into a different function that you started with.

  It's easy to get a non-zero answer for an answer that is
obviously zero by inspection. Just make the solution long
enough for the odds of making a mistake to reach 100%.