todo

波浪数值模拟

3-30-05:  write a forcing_report() function in forcing.[ch] and then link to stabilty_report 

4-4-05:  fix the GUI window stuff so that when the window is killed it kills the application.etc.
   

4-4-05:  sponge.[ch], input.[ch]  : allow for different options to sponge layer decay - linear 
           to exponential or something  ...


4-11-05:  add a _report function to give the nonlinear wave parameters, (kh, a/h  Ursell # etc)   


4-18-05: write a fancier memory report function that reports on the
         * number of field2D allocated and perhaps what types? (DONE)
         * then write a field2D stack thing so that I can allocate the temporary field2D 
           of various types on the stack... and when freed put them on the stack.
           This would save a lot of memory in all the static field2D * 
           particularly in bousinessq_dynamics.c

     
4-19-05: For the wavemaker: A number of items
	* Write a report function (DONE)
	* Fix the wavemaker depth bug - write an average water depth function at the wavemaker (DONE)
 	* Allow for wave angles, theta !=0   obliquely incident waves.   (DONE)
        * Random waves!
        * test different widths in the wave maker so that so many strange harmonics aren't generated.
        * check the spectrum at various locations downstream.....  Does it look like the correct
          type of stokes or cnoidal wave?     How long does it take for the waves to organize?


5-12-05:   General:
            * Profile the code


5-23-05: In all places where \bar_y or \delta_y is taken keep the plus info & put it into the minus for speed
      
5-24-05: In places with complicated \delta_x(), store the up column for future us in the down column
        as in the tracer advection.    This could be utilized in b_calc_eta_EP2()
        Other places?


5-25-05: NOTE that in fCshoal with theta!=0, the vorticity is <>0!  - without any wavebreaking.   This 
      on the surface seems to violate the condition of irrotationality but I think it is ok because in the
      2003 Chen et al. JGR paper  there is a lot of talk about additions to the dynamics to make vorticity conserved



5-26-05: Build the code on hull:  use valgrind to check for memory leaks and stuff!
           - use the full fCqslope1B.init file because that is the baddest mother around
      (DONE: found 4 or five really bad memory leaks and fixed most of them.  The
       one remaining is in last function in bousinessq_dynamics with the deallocation of the temporary
       field2D.   This does not remove the field2D from the linked list - minor bug )

5-27-05: Note that there is a bug if the 1-d or 2-d depth file is not the right size.  fscanf just
      ignores the error and doesn't change the value of &a.    This leads to the same bathymetry 
      continuing on I       guess but it's not good behaviour.   Need to error check.


5-29-05: there is a bug in the breaking eddy viscosity output - the ptr_to_field2D is set
     before the breaking eddy viscosity is ever set.   I have no idea how this works.


6-12-05: POTENTIAL SPEEDUPS:  * in the tridiagonal code - lots of potential prefactoring
            * what is up with threads??    can I make the code faster?

6-12-05:  simple stuff to check/fix:
             * make a better help file input
             * check the parse option.  Make sure it doesn't open any files

6-15-04: note that the wavemaker's x-width is related to the total wavenumber k not the 
     k_x component.   For waves at an angle I think this will matter.


6-24-05:  look at the wavemaker g(x), I and D in detail (this is ok)

6-25-05:  The reason the wei_kirby dynamics w/ wavemaker is so noisy is 
       the E2 term in the d(eta)/dt equation!    I checked this by taking
       out separately the  (F2,G2), (Ft, Gt) terms and then the E2 term.
       So this means either:   1) there is a bug in E2   
                               2) the wavemaker w/ wei_kirby dynamics is bad
            => actually  WK shoaling solitary wave also crashes for big enough sol wave
 

7-4-05:  The fundamental question is this:  Is the WK dynamics OK or is the wavemaker
       messed up?   Clearly the E2 terms is kinda unhappy with the wavemaker.  What
       about INITIAL CONDITION problems?   SHOLAING SOLITARY WAVE of Wei & Kirby
        - make a fCshoalsol.[init,m] and fCmshoalsol..m  files to add to CVS
           

7-4-05: Check WK w/ wavemaker results. \delta=2.  ak, kh, a/h dependence on noise?
        - is one reason wave heights vary in space because of aliasing due to saving only
          every 0.1 sec?   How about saving more often?    NOTE THAT NWOGU does not
          have this problem.   IF I turn off E2 in these runs does it fix the H(x) problem
          w/ WK in the wavemaker?


7-4-05: Make a planar beach shoaling and breaking case:  Test NWOGU & WK.  Does the
    breaking wave height follow H/h = constant????

7-4-05:  For U,V, & ETA initial conditions - use a non-numeric flag for the initial conditions

7-4-05:  Update the manual.tex, the -h flag, and other things in prep for a beta release.

7-5-05:  For the loading bathymetry.  If it says file2d but there is not enough data points return an error!!!
           - same for file1d too!!!!  ARGHHHHHH

7-7-05:  Sponge layer stuff:  make the sponge layer coefficient be independent of wave frequency         

7-12-05: RANDOM1 waves:  narrow band - ok, spectrum file loading - what is the format???  
         columns of:   freq a0 theta - so first need to know the length of this...
         * test random1file 
         * also freq-directional spectrum too  RANDOM2  !!!!
    RANDOM1: 1) data struct  2) init 3) rhs_eta_add_f...  4) deallocate  5) report 6) input
    RANDOM2: 1) data struct  2) init 3) rhs_eta_add_f...  4) deallocate  5) report 6) input
       - need a function that takes as input (theta0, spread, Ly, and # of ky) and returns a set of (ky and amplitudes)

     *** NOTE ALSO: wavemaker g_x length is still based on ktot and not k_x!!!! should be fixed
         NOTE too:   sin(a +/- b) = sin(a)cos(b) +/- cos(a)sin(b)
                     cos(a +/- b) = cos(a)cos(b) -/+ sin(a)sin(b)

7-14-05:  make a function that calculates the terms of the radiation stress tensor over a period of time
       - actually calculate wave statistics at each location???   E(f,theta) or a1, a2, b1, b2 ??


7-17-05: do a grep for "fix" and fix all the places it says fix

7-29-05: put the configure stuff into the funwaveC directory so that it can stand along even
   though it isn't alone in CVS 

7-29-05: handle the bug so that if there are carriage returns at the end of the input file there
      is no hangup in flags.c:parse_line()