Software Archeology @ RIT
Last Week
In the last few weeks we have finished the scraper and run it, to get all 159,254 code reviews in total! As Danielle said in her week 13 post, our test data (that was comprise of ~1000 files) took 60 minutes to run. When we tried to run the entire data set through the loader for the first time, it took over 24 hours.
This Week
Our efforts since the original load have been mostly focused on optimizing the loading effort so that our “rake run” doesn’t take a whole day or more. We need to have a more efficient way of processing that data.
Other things we have been working on include writing more verify tasks to ensure the integrity of our database, and making some minor fixes to our database model to affirm the relationships between the tables.
GitHub Stats
One of the most valuable tools that our team uses is GitHub, a source control website that allows you to collaborate with developers on an open source project. This system gives our team the ability to work simultaneously without collisions, and talk to one another about problems and bugs. This has been a great mode of communication for us throughout this project. Another great feature of GitHub is that it can produce interesting statistics in reference to how active your repository has been recently. Below is a graph of our commits for the life of the project.
You can see there is a large lull over summer followed by a large increase in activity at the beginning of the year, and then a slope downward at the start of intercession.
Another interesting graph to look at is what GitHub calls our “punch card”. This allows you to see the most common times that your project is being worked on. The bigger the dot, the more commits have been pushed at that time.
Here you can see that our most commons time is Thursdays at 5PM, followed by some other large dots spread across the graph.
This project is sponsored by the Collaborative Research Experiences for Undergraduates, funded by the National Science Foundation.