Go to the previous, next section.

Summary

Ingrid is a generalization of SUMRY. It, like SUMRY, is supposed to make it relatively easy to deal with the output of a model, namely generating plots and files. SUMRY is also able to read data files, plotting directly without an intervening model. Ingrid takes this one step farther, attempting to facilitate the reading in of data to models as well.

Ingrid's ability to handle both the input and output of a model means that there is a great deal of flexibility in handling the model's input and output. If one needs to create a history file so that the model can be restarted, Ingrid can both generate the file from the first run, and read the file back in for the second run. Models can also be fed idealized forcing and other specialized configurations, as well as simple functions of the input data (such as adding a mean and anomaly file before feeding it to the model, or renormalizing a data set).

Because Ingrid has a great deal of power, a command file is necessary. This is mainly because we frequently run models multiple times with slightly different parameters, different input data sets, different sorts of plots requested, and it is nice to have such a run neatly summarized in one file. Since I have a fair amount of experience with RPN languages, and they are fairly simple to implement, I am going to make it similar to FORTH/PostScript. Making it similar to PostScript reduces the amount of documentation necessary, since there are a number of published PostScript tutorials already.

The main features of Ingrid are:

Reads data from our data catalog
Reads and writes netCDF files, a self-documenting, machine-independent data file format supported by Unidata (UCAR). There are a number of public domain programs that read and write this data format. See section Preparing a netCDF File for an example of creating a netCDF file to be read by Ingrid.
Also writes HDF files, a self-documenting file used by many NCSA programs as input files.
Attaches to a model, such that it can both supply data to the model and extract data from the model.
Generates PostScript plots, including line, contour, vector, and scatter plots, as well as histograms.
Plots and files are generated as soon as the data is available, so that Ingrid can be used for debugging: all plots of the model before the point where the model fails are available.

Go to the next section.