I've collected links to free, open source software apps used in several countries for COVID contact tracking jgbarah.github.io/Notes/foss-cov… Including links to source code *and* Cauldron.io metrics visualization, because transparency matters. Thread ⬇️⬇️⬇️ 
OpenTrace (Singapore) was maybe the first one to be released. To some extent it is more an upstream for others to borrow than the actual app. It was more like dumped than developed in the open, and seems stagnant since May cauldron.io/dashboard/1566…
With only 39 commits, OpenTrace was likely developed elsewhere, with only close to final code published. Just a few issues, and even less pull requests, seem to validate that impresion. But others followed...
CoronaWarnApp (Germany) is one of the most transparent. Commits are available since late April, and the size of the developing team, the commit count, and the number of issues & pull requests suggest the action is happening in the open cauldron.io/dashboard/1545…
Reviews for CoronaWarnApp (in the form of pull requests) are solved *quickly* (1.12 days), and numbers suggest that a large fraction of commits go through code review. You can also see pull requests being opened and closed even during weekends, suggesting high engagement.
Commit and churn patterns for CoronaWarnApp suggest a large refactoring in early June, close to its release. Which implies both preparation for the release, and attention to the feedback of early usage
For the case of CoronaWarnApp, developed by companies, it is no surprise that most of the action happens in office hours. But see those commits after eight, and that high activity during weekends. The size of the development team is large: ~100 persons
Stopp Corona (Austria) also seems to be developed in the open, with a history of ~850 commits since late April, but in this case with a team of about 10 persons cauldron.io/dashboard/1559…
Stopp Corona activity in code review and issue tracking matches that of commits, with quick reviews (<1 day), but slower attention to issues (~16 days). However, most of the issues backlog seems to come from release time: since early May, closed issues almost match opened issues
Activity for Stopp Corona clearly matches that of office hours: almost no activity in commits, or opening or closing issues during weekends, almost all commits from 9:00 to 19:00
AarogyaSetu (India) was published in a burst of about 100 commits during late May, and has not been touched since then. So, nothing to say about its development process, except that the source code in its public repos has not been updated during the last two months.
Inmuni App (Italy) shows its development history since mid-March (one of the longest histories) cauldron.io/dashboard/1561… Its activity pattern suggests it was developed internally for ~2.5 months, then released, with a spike due to accommodation to real usage
For Inmuni App, all visualizations for lines added/removed, issues opened/closed, and prs open/closed have very similar patterns since mid May, suggesting these activities are coordinated.
Inmuni App has also a clear corporate activity pattern, although some activity off-hours can be found. The fact that opening issues & pull requests follow the same pattern suggest it is the development team mainly performing that activity.
Inmuni App maintenance team seems to be currently of about 5 persons, with peeks of 15-20 persons per week working on it.
NHS COVID-19 (UK) has a rich history, with more than 3,600 commits since mid March. Looking at the volume and the patterns, it is one of the most transparent projects, with mostly real activity in its repos cauldron.io/dashboard/1565…
However, for NHS COVID-19, there is really little action in issues and pull requests, which suggest that they are doing issue tracking and code review elsewhere. So, we don't have that info.
Since late June there is a complete black-out in commits. It seems the development was stopped, or completely moved elsewhere. Curiously enough, this was following an extensive clean-up of the code, with deletion of many lines (preparing for shutting-down?)
eRouška (Czech Rep.) has also a wealth of information in its repos cauldron.io/dashboard/1577… Most of the activity is from mid-March to mid-May. Did the apps reach maturity about then?
During peak time, about 10 persons were active committers per week. Number of pull requests suggest they are mostly using them for code review, with a review time of about 1.5 days.
For eRouška (Czech Rep.), looking at lines added/removed,most of the work was done by mid-April. And the hourly and daily patterns suggest a richer community, not only companies working on it: a lot of activity late in the evening, and during weekends
Hamagen (Israel) has quite a constant activity in git commits since mid-March up to now cauldron.io/dashboard/1578… The team seems small (~10 committers per week at peaks, <5 most of the time)
Except for a short time at the beginning of its development history, issues and pull requests for Hamagen apps are just a few, which suggest review and issue tracking are happening elsewhere
However, activity in the source of Hamagen code seems very intense with at least one refeactoring in mid-May, which included likely new functionality.
Hourly and daily working patterns for Hamagen seem clearly those of developers hired by companies, working in Israel (notice that Fri-Sat weekend). Most of the development is 10:00-18:00.
COVID Tracker (Ireland), released recently, has a very short history, suggesting it was developed somewhere else, and then dumped in its public repos cauldron.io/dashboard/1766… Therefore, nothing to say about it, for now.
This is all for now wrt the development history of COVID contact tracking apps. If you know some others, please let me know, I'd love to include them in the analysis, and have a cauldron.io dashboard for them
All in all, I'd say that most of the apps are quite transparent. Of course, maybe a part of the development is hidden, but for many of them patterns seem legit, showing everything. Proxy developers, commit splitting and other practices could be happening, but that's not likely.
From my personal pov, the fact that development history for these apps is shown in public, in addition to them being free, open source software, is a new step in transparency that we all should appreciate and request for the software we use speakerdeck.com/jgbarah/open-d…
