                 Known problems 

  o Very occassionally (about 1 in 1000 in monte carlo experiments) the
     algorithm gets stuck and does not improve (so does not converge). The
     workaround is to restart with a very slightly perturbed starting point.       

            Changes in GPArotation version 2008.05-1

  o added echelon rotation.

  o added gradient Gq to result list from  GPForth and GPFoblq.

  o change license from "GPL-2" to "GPL-2 or later".

 
            Changes in GPArotation version 2007.06-1

  o fixed a couple of lingering $Lh (in print and summary methods) that should
    have been changed to $loadings.


            Changes in GPArotation version 2007.04-1

  o removed an extra comma in c() that caused a test failure with R-2.5.0

  o added eiv rotation.
    
  o renamed $Lh in the result from GPForth and GPFoblq to $loadings. As a
    result, rotation methods calling these function no longer need to rename
    this element in order to work with factanal and other programs. (It is a
    good idea to use the extractor function loadings() rather than refer
    directly to object structure.)

  o changed rotation method functions to return all elements of GPFoblq and
    GFForth.

  o fixed the documentation file primary alias for all rotations (which was 
     being called oblimin).
      

            Changes in GPArotation version 2006.2-2

  o extra argument (...) to invisible in print.GPArotation was removed.
  
            Changes in GPArotation version 2006.2-1

  o broken references in documentation were fixed and updated.


            Changes in GPArotation version 2005.10-1

  o warning message about non-convergence expanded to indicate function.  
  

                  GPArotation version 2005.4-1

  o  First released version.
