Draft of JOSS paper.

Also, remove defunct LGO directory.

.zenodo.json

@@ -1,5 +1,5 @@
 {
-    "description": "RK-Opt: Optimization of Runge-Kutta and related time integration methods",
+    "description": "RK-Opt: A Package for the Design of Numerical ODE Solvers",
     "license": "bsd-license",
     "title": "ketch/RK-Opt",
     "upload_type": "software",
@@ -21,7 +21,7 @@
         },
         {
             "affiliation": "Capital One",
-            "name": "aron@ahmadia.net",
+            "name": "Aron Ahmadia",
             "orcid": "0000-0002-2573-2481"
         }
     ],

README.md

@@ -1,29 +1,34 @@
 
-# RK-opt
+# RK-opt: A Package for the Design of Numerical ODE solvers
 
 See the full documentation [here](http://numerics.kaust.edu.sa/RK-opt).
 
-RK-opt is a collection of MATLAB code for designing optimized Runge-Kutta methods.
+RK-opt is a collection of MATLAB codes for designing optimized numerical ODE solvers.
+The main emphasis is on Runge-Kutta methods, but some routines deal with other classes of methods.
 It is primarily developed and used by the 
 [KAUST Numerical Mathematics Group](http://numerics.kaust.edu.sa).
 It includes the following sub-packages:
 
+ - **polyopt**: Find optimal stability polynomials of a given degree and order of
+   accuracy for a specified spectrum.
  - **RK-coeff-opt**: Find optimal Runge-Kutta method coefficients, for a prescribed
    order of accuracy and number of stages.
  - **am_rad-opt**: Find stability functions with optimal radius of absolute monotonicity.
    Includes capabilities for both multistep and multistage methods.
- - **polyopt**: Find optimal stability polynomials of a given degree and order of
-   accuracy for a specified spectrum.
  - **RKtools**: A collection of routines for analyzing or computing various 
    properties of Runge-Kutta methods.  For a much more extensive package along these
    lines, see [NodePy](http://numerics.kaust.edu.sa/nodepy).
 
+A common workflow for designing Runge-Kutta methods is to use **polyopt** to find an
+appropriate stability function and then **RK-coeff-opt** to determine the Runge-Kutta
+method coefficients.
 
 # Authors
-The following people have contributed code to RK-opt (listed alphabetically):
+The code is primarily developed and maintained by David Ketcheson.
+The following people have also made important constributions to RK-opt (listed alphabetically):
 
- - Aron Ahmadia: **polyopt** routines
- - David Ketcheson: Principal author and maintainer
- - Matteo Parsani: **RK-coeff-opt** routines and organization
+ - Aron Ahmadia: Co-developer of **polyopt** algorithm and routines.
  - Zachary Grant: Extension of order conditions to multistep RK with more than two stages and 
     addition of order conditions for orders 9-11.
+ - Matteo Parsani: Many improvements to **RK-coeff-opt** routines and organization.
+ - Hendrik Ranocha: General improvements and updates, including updating the test routines.