The development strategy behind Arcan has always been to work with experimental proof of concepts doing ‘traditional’ tasks in odd way and use that as feedback to refactor and improve the Engine, API, testing and documentation.
For instance — the Video decoding, encoding and tagging experiments added process separation and greatly helped shape the shared memory interface. The arcade frontend experiments à Gridle improved support for odd input combinations (2 mice, 3 keyboards and 5 gamepads? not a problem), support for synchronization with time sensitive processes (libretro frameserver) where buffering and other common solutions were not available. The AWB experiments helped define controlled and segmented data sharing, along with performance considerations in tricky UI situations (hierarchies of windows where size and position relied on dynamic sources, drag+resize and watch hell break lose).
The end goals, getting a portable graphics- focused backend for putting together embedded, mobile and desktop system interfaces; balancing security, performance, stability and no-nonsense style- ease of use — is still out of reach, but great strides have been made. The last couple of months have mostly been stuck documenting, testing and working with the corners that dynamic multiscreen entails.
The next experiment in this regard is Senseye, which is targeted towards the more rugged of computing travellers; the reverse engineers, the security ‘enthusiasts’ and the system analysts. It is a tool for navigating and controlling non-native representations of large, unknown, binary data streams and blocks. Both statically in terms of files and dumps, and dynamic through live acquisition of memory from running processes.
The video above shows using senseye for navigating a suspicious binary and for poking around the memory pages allocated by pid 1 (still init though in its twilight..).