Data collection for this study was performed by
the Github project this page belongs to.
The entry point to the analysis is the Bourne shell script
ifdefrevolve-project.sh, located in
ifdefrevolve-project.sh -h at the command line to
get a help screen explaining the command line options.
This script relies on a number of helper scripts, which reside in
the same directory.
To get everything to run, this directory should be in the
variable in your environment.
More information can be found on the tool support page.
Functions can change from one snapshot to the next. For example, functions are modified (becoming longer or shorter), and new functions are added, existing ones are deleted or moved to new files. All those changes invalidate (to some extent) the static metrics, which we collect only once per snapshot. We also lose some information about functions being added later on or moved to new files also.
To estimate how small we need to make our snapshots in order to get
accurate results and not miss too much information, we conducted an
initial experiment on
OpenLDAP. On a separate page, we list the
results of the snapshot size experiment.
The data for all eight subject systems used in the study is
available here. The data is
available both in
.csv format and as R data sets (files named
The names and repository URLs of the subject systems are listed below.
|Subject Name||Repository URL|