Some requests
Posted: Fri May 01, 2009 7:59 am
I use Eviews a lot and think that it is great. Nonetheless, there are a number of additions that I (and the others I work with) would find useful.
Programs
1. Have error messages that are generated during the execution of a program indicate the line of program or subroutine code in which the error occurred.
2. The ability to specify the maximum number of permitted errors within a program or in a command that executes a program.
Models
3. A command to "drop" an equation that has been previously appended to a model (so that an alternative equation can be appended or merged in its place)
4. Generalize the "control" proc to allow for multiple pairs of endogenous and exogenous variables to be switched.
5. Improve the "solve" proc so that when a "solver has stalled" error is encountered, the solution algorithm searches in a different direction or shifts to a different algorithm.
Matrices and vectors
6. Commands or functions that create a new matrix (vector) from an existing matrix (vector) based on extracting or eliminating sets of non-contiguous rows/columns (elements), as given perhaps by the contents of a vector.
7. LU and QR matrix factorization functions.
Programs
1. Have error messages that are generated during the execution of a program indicate the line of program or subroutine code in which the error occurred.
2. The ability to specify the maximum number of permitted errors within a program or in a command that executes a program.
Models
3. A command to "drop" an equation that has been previously appended to a model (so that an alternative equation can be appended or merged in its place)
4. Generalize the "control" proc to allow for multiple pairs of endogenous and exogenous variables to be switched.
5. Improve the "solve" proc so that when a "solver has stalled" error is encountered, the solution algorithm searches in a different direction or shifts to a different algorithm.
Matrices and vectors
6. Commands or functions that create a new matrix (vector) from an existing matrix (vector) based on extracting or eliminating sets of non-contiguous rows/columns (elements), as given perhaps by the contents of a vector.
7. LU and QR matrix factorization functions.