Will Neostring ever be released on Android?

[vairysobik]

Neostring is a really cool extension; I'm currently implementing it in my PC-Only game. But as far as I know, Neostring doesn't currently work on Android. Will there ever be an Android port?

[defisym]

In short, porting to other platforms has two major challenges:

1. Originally, the project was just a wrapper for Fusion's text rendering engine, which was based on GDI under DirectX 9, then move to native GDI/GDI+ later due to performance bottlenecks. Since a portion of the codebase already existed at that time, the entire rendering system was built around GDI/GDI+. Furthermore, because development was driven strictly by the immediate needs of an active game project, the implementation of the format parsing logic is quite crude and lacks proper abstraction (with many nested lambdas tangled together). To port this to other platforms, a complete refactoring is required first to encapsulate all GDI-related code, followed by implementing a new solution based on alternative font rendering libraries. This would demand a significant investment of time and effort.

2. Fusion's inherent design is fundamentally problematic. While its cross-platform approach shares assets and events, it completely rewrites the runtime for each platform using entirely different programming languages; using native approaches is practically equivalent to developing from scratch. Although C/C++ is compatible with almost all platforms, the documentation for Android NDK is quite poor (in fact, documentation is lacking everywhere). Even if reuse the same C/C++ codebase, it will take a massive amount of time to adapt—such as figuring out how to submit rendering results for display. Fusion lacks an internal cross-platform abstraction, such as a unified rendering interface. The time spent adapting each platform individually might even exceed rewriting it from scratch in Java. However, doing so would create a new maintenance burden, aka manage multiple codebases for any future feature updates.

Therefore, regarding cross-platform support, IMHO the most rational decision is to wait for the brand-new SDK provided by Fusion 3. Since we will have to re-adapt the codebase anyway, it makes more sense to wait and solve the problem once and for all.