Abstract
This thesis comparatively analyses decentralized social media platforms' architectures while developing a taxonomy to classify their distinct characteristics focusing on privacy, identity management, governance and moderation. Through detailed examination of decentralized and federated platforms such as Mastodon, Lens, Bluesky and others, this thesis highlights the similarities and differences of these platforms' design choices. While examining the trade-offs of each model, it highlights that no single architecture is optimal, instead these design choices allow them to have unique strengths that align with platform's goals and user base. The platforms and findings from the comparative research used to develop a systematic taxonomy using an iterative approach to demonstrate how these platforms can be grouped and reveal characteristics more clearly. The results of the comparative analysis and taxonomy development are then combined to suggest key design characteristics for a user oriented distributed news platform. This example design attempts to find a balance between giving users freedom while ensuring sure that moderation is effective and information flows reliably by using a hybrid architectural approach. This thesis contributes to the broader understanding of how distributed architectures can contribute to the digital social platforms in terms of privacy, identity and governance.
Research Questions
1 - How do centralized, federated, and decentralized social media platforms compare in addressing technical architecture aspects such as privacy, identity management, and governance?
2 -What are the functional characteristics of distributed social media platforms, and how can these differences be systematically categorized using Nickerson’s method for taxonomy development?
3 - What design parameters are essential for a user oriented distributed news platform?
Name | Type | Size | Last Modification | Last Editor |
---|---|---|---|---|
kick-off-presentation.pdf | 1,87 MB | 27.05.2025 | ||
Mehmet Efe MT Final.pdf | 1,62 MB | 27.05.2025 |