(Super) Apps evolution
: Jason Chao, Michael Dieter, Anne Helmond, Esther Weltevrede
: Andrea Benedetti (DensityDesign)
In this sub-project we employed AppInspect to examine app packages over time to understand app evolution. In particular, we explored how super apps and other apps evolve on the technical level of the app software package (Dieter et al., 2019
). Despite being dynamic objects which continuously update themselves, we can use app repositories to collect older app packages and analyse them over time in order to write app histories (Helmond and van der Vlist, 2021
3. Research Questions
How do apps evolve over time in terms of their infrastructural dependencies?
4. Methodology and initial datasets
We collected old software files for three super apps and for TikTok, from both international and local APK repositories. The timespan of our data is ten years.
We found that most super apps grow in filesize over time. VK and WeChat have the same file size over time, indicating that the international and local version the same (which Jason Chao later technically confirmed). TikTok ’s international and local versions however do clearly show different sizes - and are confirmed two different versions.
Regarding permissions, we also find a rise in the number of permissions requested over time. An open question is whether this relates to increasing functionalities and mini-apps within those apps. Here again we see a significant difference between the international and Chinese TikTok. The Chinese version requesting a staggering amount of 230 permissions.
Over time, we also see more sensors invoked.
Next, regarding trackers we see various patterns emerge. Both VK and WeChat contain their own parent company’s trackers as well as many Google and Facebook trackers.
TikTok international also makes use of many Google and Facebook trackers as well as a few Chinese trackers - while the Chinese version mainly contains Chinese trackers - which expand over time. Interestingly, in 2019, TikTok paid $5.7 million to the Federal Trade Commission for violations of America's children's privacy law. Afterwards we see that TikTok has removed its Chinese trackers from its international version, explaining the decrease in trackers!
The discussion concerns our methodological reflections as these were our first ventures into studying app software evolution. We have found multiple challenges for collecting APKs over time:
- Different versioning, alignment with dates. There’s no absolute list of app versions and dates.
- There may be gaps in data because some apps cannot be unpacked due to encryption.
- The importance of static versus dynamic app analysis. E.g. very clear with super apps which instantiate themselves and their functionalities based on user’s (store geo) location.
Helmond, Anne and Fernando N. van der Vlist (2021) Platform and App Histories: Assessing Source Availability in Web Archives and App Repositories. In: Gomes D, Demidova E, Winters J, et al. (eds) The Past Web: Exploring Web Archives
. Cham: Springer International Publishing, pp. 203–214. DOI: 10.1007/978-3-030-63291-5_16
. author pre-print PDF
Dieter, Michael, Carolin Gerlitz, Anne Helmond, Nathaniel Tkacz, Fernando N. van der Vlist, and Esther Weltevrede (2019). Multi-Situated App Studies: Methods and Propositions. Social Media + Society
, 5(2), 1–15. https://doi.org/10.1177/2056305119846486