Project Everest began at Microsoft Research in 2016, aiming to spur research in program verification to produce industrial-grade software. In collaboration with INRIA and Carnegie Mellon University, Project Everest’s goal was to produce drop-in verified replacements of secure communications software used in the HTTPS ecosystem, including TLS, the underlying cryptography, and related subprotocols. Now, almost a decade later, we reflect on the project, sharing both its successes and failures, and look ahead to the next decade of program verification research.

Library OS (LibOS) architectures implement the OS personality as a user-mode library, giving each application the flexibility to choose its LibOS. This approach is appealing for many reasons, not least the ability to extend or customise the LibOS. Recent work with Drawbridge [29] showed that an existing commodity OS (Windows 7) could be refactored to produce a LibOS while retaining application compatibility.

This paper describes the techniques used to describe and synthesize FPGA circuits expressed in a data-parallel domain specific language (DSL) called Accelerator. We identify the subset of data-parallel descriptions that are supported by our system and explain how we track memory access patterns which allow us to generate efficient FPGA circuits.

We introduce activity-based navigation, which uses human activities derived from sensor data to help people navigate, in particular to retrace a "trail" previously taken by that person or another person. Such trails may include step counts, walking up/down stairs or taking elevators, compass directions, and photos taken along a user's path, in addition to absolute positioning (GPS and maps) when available. To explore the user experience of activity-based navigation, we built Greenfield, a mobile device interface for finding a car. We conducted a ten participant user study comparing users' ability to find cars across three different presentations of activity-based information as well as verbal instructions. Our results show that activity-based navigation can be used for car finding and suggest its promise more generally for supporting navigation tasks. We present lessons for future activity-based navigation interfaces, and motivate further work in this space, particularly in the area of robust activity inference.