Cookbook

1: Introduction
2: Invoking Gre
3: Gre Grammar
4: Commands
5: Cookbook
6: Hints
7: History
8: Plans
9: Concept Index

5: Cookbook

5.1: Example 1 -- filled-curves, dots, line-segments

-- to be filled in later --

5.2: Example 2 -- curves, arrows, labels

Diagram illustrating x-y curves, arrows, text. Since this is a definition sketch, no axes are drawn. (From Kelley and Van Scoy, 1997)

5.3: Example 3 -- color-coded dots, isopycnals

Diagram illustrating temperature-salinity plot, color-coded for a third variable (in this case, bug counts). Data from bioness tows made by Dalhousie as part of the OPEN program.

5.4: Example 4 -- a cmd for TS diagrams

The `cmd' created in this example draws TS diagrams, setting up axis names and drawing isopycnals automatically.

5.5: Example 5 -- timing test

The following code tests the speed of programming a subroutine in gre, compared to using a builtin function. To run it, do e.g. `time gre test.gre', then change the comment in the lines inside the loop, to use either builtin code or the user-defined function.

Depending on the speed of your machine, you may have to use a different value of `$n', in order to get times long enough to compare, using the unix `time' program. Also, you may wish to duplicate the function call a few times, to see whether the loop overhead is significant.

My results, on a Sun Microsystems solaris machine, indicate that the gre-based `my_sinh' subroutine is 3.6 +/- 0.1 times slower than the internally coded `sinh' subroutine. Because sinh is a slow function to calculate, the loop overhead is negligible -- that is, increasing the number of calls to the subroutine inside the loop changed results very little. 

   # Test of timing
   sub my_sinh ($) {  # hyperbolic sin
       my $y;
       $y = exp($_[0]);
       return ($y - 1 / $y) / 2;
   }
   $n = 10000;
   $dx = 1 / $n;
   for ($x = 0; $x < 1; $x += $dx) {
       # Uncomment one of following two lines
       #$sy = my_sinh($x);
       $sy = sinh($x);
   }

5.6: Example 6 -- plotting theoretical functions

Diagram from my introductory Physical Oceanography class, illustrating (a) large fonts and pretty colors, suitable for overhead projection and (b) display of theoretical curves as oppposed to data.

5.7: Example 7 -- drawing a coastline map with fancy projection

Diagram shows Maritimes provinces of Canada, plotted using the `set map projection' command with Lambert Conformal Conic projection see Set map projection.

5.8: Example 8 -- colored text

Face it, people like/hate to be reminded of their age on birthdays. So, the way to make them like/hate you is to wrap their present in paper that was generated as below. Only a snippet of the 8.5x11" output is shown. Click on the diagram to see the source code, and just change the `$year = ...' line to customize for your friends.

5.9: Example 9 -- month-of-year timeseries

There is no builtin facility for plotting against month-of-the year, but you can get this sort of graph pretty easily. For daily data you'll need to know the number of days in each month (as a function of whether it's a leap year or not), etc. But for month-averaged data, the case is simpler; you just need to write `Jan', `Feb', etc, along the x-axis. The following example assumes that the January data will have x-value of 0.5, the February data will have 1.5, etc.

(c) 1997-2000 Dan E. Kelley, email Dan.Kelley@Dal.Ca