Answer to RQ2 and RQ3

[akondrahman]

Next steps:

• RQ2: need to identify raw occurrences, smell density, and proportion of scripts with smells
• RQ3: label scripts automatically so that each script is labeled '1' if it contains at least one smell, otherwise 0. Then mine metrics (mining script previously provided)

Need it by May 05.

[rayhanur-rahman]

RQ2: https://github.com/brokenquark/SLIC-Ansible/blob/master/occurence-density-proportion.csv RQ3: https://github.com/brokenquark/SLIC-Ansible/blob/master/V1_TEST_AUTHOR_ANSIBLE-2.csv Ansible Smell Count in two different formats: https://github.com/brokenquark/SLIC-Ansible/blob/master/ansible-smell-count.csv https://github.com/brokenquark/SLIC-Ansible/blob/master/ansible-smell-count-different-format.csv

[akondrahman]

Label files properly ... 'occurrence-density-proportion' for both GitHub and Openstack repos.

[akondrahman]

Combined smell density calculation is wrong ... it should be

(total smell count in all files / (total lines of code in all files / 1000) )

[akondrahman]

Also double check proportion calculation. This is the formula:

((no. of files with at least one dataset) / (all files in dataset)) * 100

[rayhanur-rahman]

Link: https://github.com/brokenquark/SLIC-Ansible/blob/master/RQ2/data.txt

[akondrahman]

New tasks:

• Need RQ3 results for Openstack .... I only see GitHub
• Write a paragraph of 4-5 sentences on why security smell frequency is 10x lower for Openstack than GithUb.

[rayhanur-rahman]

It will take time, because I never downloaded openstack repo. Just parsed the ymls. Hopefully tonight.

[rayhanur-rahman]

• Write a paragraph of 4-5 sentences on why security smell frequency is 10x lower for Openstack than GithUb.

I have fixed the density calculation. There was a silly mistake. Do I need to explain why openstack has fewer occurance? Now they are reasonably comparable with github.

RQ3 for openstack is available now