User Interface Architecture
Some Design Patterns
Security of X.org
Let the API freeze by waiting, till a lot of popular applications depend on the API. After that re-implement the project by testing the implementation with formal methods and possibly also with other means. Applying speed optimizations is part of the job. The re-implementation will probably have to be dynamically configurable, because the API of the initial implementation probably introduces fundamental security flaws. Those flaws have to be sand-boxed in the new implementation, so that one application is not able to abuse/infect other applications or lacks the functionality that depends on the flawed API or the flawed API is intentionally made to behave some non-bakcward-compatible way that does not crash the applications that use the flawed API.
Re-implementing abandonware like the Windows API is a waste of time. Re-implementing the sloppy technology that is slammed together for business software development is a waste of time. The C# virtual machine, Mono, might be re-implemented, only after the Microsoft has totally lost interest to it, no corporate crap-ware developer has any monetary incentive to introduce any sloppy, dirty, changes to it and no CEO thinks that it can use the project for extorting money from someone, like it is with the Oracle(Larry Ellison, archival copy) and Java in 2016.
- File:2009 02 xx Securing Graphical User Interfaces by Norman Feske.pdf
- File:A Nitpicker s Guide to a Minimal-complexity Secure GUI by Norman Feske and Christian Helmuth.pdf
- Raudrohi State Cluster Specification Type 1 (arcieval copy)
- File:Generating Commercial web Applications from Precise Requirements and Formal Specifications David Crocker and John H Warren.pdf