A simple app to extract stats about repositories using the Github API and Cloc
- Github: The repo information is completly gathered from Github. There are a number of libraries you can use to accomplish this. We are using @octokit/rest.
- To use the Github API more efficiently you should use a Personal Github Token. You can then place this token in a
.envfile at the root of the project. For convinience this repo includes a .env-example file that you can rename and use.
- Cloc: Part of the analysis we are conducting is the cost of software. We are using Cloc to get a project's lines of code. This total is then used with COCOMO II to calculate costs and savings. You can find information for CLOC here.
- Releases.json: the Code.gov front-end uses a file called releases.json where all our code inventory is stored for use. This is going to change in the near future for the use of our API.
Code.gov estimates the evaluation of the federal open source program by using the Constructive Cost Model II -- CoCoMoII.
- COCOMOII uses a regression formula of historical projects to estimates, size, effort, duration and cost of the software.
- Code.gove uses COCOMOII by counting lines of code in a repository, and assigns an hourly rate of $48 -- the median hourly salary for a software developer and programmer according to the Bureau of Labor and Statistics.
- Since CoCoMoII calculates based on person-months (19 days), we adjust $48/hr to a monthly cost -- $48(19 days x 8 hours) = $7,296. We keep all factors, such as Risk and Analyst Capability as nominal.
- COCOMOII generates an estimated cost to create the software and the amount of time in person-months to develop the program.
- The cost becomes our estimated cost to the government to create, or investment.
- Code.gov multiplies cost by the number of agencies using the program, omitting the creating agency.
- This becomes our estimated cost-savings, or estimated net profit
- Using investment and net profit we are able to calculate the government’s estimated return on investment over the project lifespan and annualized by taking an estimated yearly average.
- To calculate labor hours, we take the person-months and multiply by 152 -- (19 days x 8 hours).