Google Code-in 2017 is over. You can read the wrap-up to see who won for our organization, while this blog posts summarizes statistics on the work done during the program.
Period
The program ran from 28 November 2017 to 17 January 2018, for a period of 51 days, during 7 weeks, see the timeline.
Mentors
We had 12 mentors involved in the program, listed in the order of comment interactions with the tasks inside the Code-in application:
- Thomas Mortagne, Alex Cotiugă, Eduard Moraru, Ecaterina Moraru - interacted with more than 60 task instances
- Pawan Pal, Vincent Massol, Shubham Jain, Clément Aubin - interacted with more than 20 task instances
- Marius Dumitru Florea, Cristina Rosu, Guillaume Delhumeau, Denis Gervalle - interacted with less than 10 task instances.
There were also 1094 comments between mentors and students on the opened code Pull Requests, plus a lot of discussions inside our #xwiki IRC channel.
Mentors estimated to have spent around an average of 1.5+ hours per day (average of 4+ hours per week) to manage the tasks, pull requests and communications with the students. There was additional time needed to provide / sort the Onboarding issues, adding them to the Code-in application, computing the GCI statistics, etc.
Students
We had 756 students that interacted with the XWiki organization in one way or another on the Code-in application. Also compared to the same time frame last year, we had a 13.24% increase in the number of visitors on xwiki.org.
We had 380 (50.3%) students completing at least 1 task. From these students, we have 34 (8.9%) students that did more than 1 task and we have 23 (6.05%) students that committed usable code in our repositories.
So 376 (49.7%) students couldn't complete any task. We have 7 (1.9%) students that tried and failed 2 consecutive tasks.
Tasks
We had 86 initial available tasks, from which 32 tasks were repeatable tasks that could be claimed by multiple students, while the rest of 54 were unique tasks. Also 40 of these tasks were marked as Beginner tasks, so we considered 47.1% of the tasks to be easily implemented by anyone.
The 756 students generated 928 task instances. From the 928 task instances, we have 439 abandoned tasks, which means a 47.3% abandoned rate for tasks. Since we have such a great number of abandoned tasks, would be nice to receive some feedback on the perceived difficulty of the tasks.
From the 489 (52.7%) successful completed tasks, the "Become a community member" has a 72.2% participation rate with 353 instances.
Our top 5 tasks completed are:
- Become a community member 353 instances (72.2%)
- Setup development environment 38 instances (7.8%)
- Fix 1 coding violation 14 instances (2.9%)
- Improve 5 English text entries 10 instances (2.0%)
- Find 5 false positives in coding violations 8 instances (1.6%)
Only 1 student updated XWiki.org screenshots and only 1 student created new test cases for test.xwiki.org.
Pull Requests
The GCI students opened 237 Pull Requests (PR), 38.8% for XWiki Platform and 50.6% for XWiki Rendering.
From the 237 PRs, we merged 65% (154 PRs) into our code base. While the majority were codestyle changes fixing coding violations (60.4%), fixing typos and rephrasing translations (21.4%), updating the XAR format (11%), there were also 11 issues (6.5%) fixed that are now part of our 9.11 and 10.0 releases:
We had 33 PRs (21.4%) fixing translations for 6 Contrib applications (Blog, Forum, GitHub, Hubot, Jira and LDAP) and XWiki Platform. There were 181 translations improved, either by fixing typos, fixing grammar or rephrasing.
Contributors
The following 23 students (sorted by contributions) committed code in our repositories:
It would be great to see some of these students stick around with XWiki community over the coming years, as mentors, as contributors or even as promoters. We loved having you around us and it would be amazing to work together again.
We thank all the students for all their contributions. It was a great Google Code-in 2017!
