From: fateman@cs.berkeley.edu   
      
      If you really want to know a symbolic form for this   
   integral, approximately, with possible relevance to some definite interval,   
     you can expand the integrand into a taylor series   
   about some suitable nearby point in that interval, and integrate that.   
      
   Using this taylor / integrate  procedure you can also get a check on whether   
   your answer is right, by computing a taylor series of the answer.  Ignore   
   the constant terms though.   
      
   This can actually be much improved over other checking techniques since   
   simplification of taylor series results may be more straightforward than   
   mucking about with some random collection of algebraic, transcendental,   
   and special functions.  I found that I had to replace abs(x) with x in some   
   cases.   
      
   This does not so much address the question of integrability in finite terms   
   as finding an integrable representation of something approximating the   
   integrand.  Which might be a suitable replacement of a possibly   
   unsolvable problem with a useful substitute.   
   It certainly works for the problem below.   
   RJF   
      
      
   On 3/16/2018 9:30 AM, Nasser M. Abbasi wrote:   
   > On 3/16/2018 10:18 AM, clicliclic@freenet.de wrote:   
      
   ...snip...   
   >>   
   >> integrate((5*x-9*sqrt(6)+26)/((x^2-4*x-50)*sqrt(x^3-30*x-56)), x)   
   >>...snip...   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)