Comment Smart Sorting
Init Manga v2.4.1 adds a new Comment Smart Sorting option in the Discussion settings, allowing administrators to control how comments are displayed. Instead of relying solely on time-based ordering, the system now supports multiple modes such as most liked, least liked, and statistically ranked comments.
This flexibility makes it suitable for both casual engagement and deeper discussions.
Wilson Score – Confidence-Based Ranking
At the core of this update is the Best (Wilson Score) mode. This statistical method ranks comments based on both the like/dislike ratio and the confidence level of the data.
Unlike traditional sorting that favors raw like counts, Wilson Score prevents low-sample comments from dominating while ensuring that highly engaged and reliable comments rise to the top.
Like & Dislike Reaction System
To support the new ranking model, the reaction system has been upgraded from a single like button to a dual like/dislike system. Users can toggle between the two states, ensuring that each comment maintains a single, clear reaction.
This creates a more balanced dataset and improves the accuracy of comment evaluation without adding friction to user interaction.
Backward Compatibility & Extensibility
The update preserves full backward compatibility by extending the existing like-comment endpoint to handle both like and dislike actions. No new endpoints are required, and existing integrations continue to function without modification.
Additionally, new hooks such as init_manga_wilson_z and init_manga_wilson_score provide deep customization capabilities for advanced use cases.
Interaction-Aware and Role-Based Ranking
The ranking system incorporates interaction-aware weighting using log10(n+1), ensuring that comments with higher engagement are treated as more reliable.
It also introduces role-based prioritization: post authors receive strong visibility for their comments, ensuring official responses are easy to find, while VIP users receive a subtle boost that enhances visibility without disrupting overall balance.
Performance Optimization
To maintain performance at scale, the system applies lightweight static caching during sorting, reducing redundant database queries when processing large comment sets. This ensures consistent speed and responsiveness even on high-traffic pages.
Conclusion
With version 2.4.1, Init Manga transforms its comment system into a fully-fledged ranking engine.
By combining statistical scoring, user interaction signals, and role-based weighting, the platform delivers a smarter, fairer, and more engaging discussion experience while remaining simple to integrate and extend.
Comments