Add small example.

1 files changed, 33 insertions, 0 deletions

diff --git a/README b/README
new file mode 100644
index 0000000..3abde32
--- /dev/null
+++ b/README
@@ -0,0 +1,33 @@
+This is a simple, quick-n-dirty wrapper for plotting .csv files with
+Gnuplot.
+
+The tool will remove comments (lines starting with a '#' character) before
+passing the data to Gnuplot. This makes it possible to have experimental
+setup and date information encoded in the .csv file, which is beneficial
+for archival reasons.
+
+You can plot the example file to a PNG graph using the following command:
+
+./gnuplot.py -o example -f png --xlabel "Time" --ylabel "Supply Bound" \
+     example.csv 1 2 "Supply Bound Function"
+
+
+In general, each curve that you want to plot in a graph is defined by four
+arguments:
+
+./gnuplot.py [OPTIONS] \
+   <file1> <xcolumn1> <ycolumn1> <title1> \
+   <file2> <xcolumn2> <ycolumn2> <title2> \
+   ... and so on ...
+
+For large experiments with many graphs, it is more convenient to wrap gnuplot.py
+in a project-specific plotting script. Have a look at plot.py and plot_ohead.py
+to see how I did it. Basically, the scrips infer what to plot and how to name
+and label the graphs from the name of each .csv file. 
+
+For example, the file name 'hard_rtss09_dist=uni_medium_250.csv' encodes the
+following information: 1) it uses hard real-time schedulability tests, it was
+using the RTSS'09 experimental setup and overheads, the task sets were generated
+with uniform utilization from the range [0.1, 0.4] (medium), and periods were
+distributed uniformly in [50, 250]. This is parsed by plot.py, which then 
+calls gnuplot.py to generate a graph with the correct labels and settings.