IMPORTANCE In patients who undergo mechanical ventilation during surgery, the ideal tidal volume is unclear.OBJECTIVE To determine whether low-tidal-volume ventilation compared with conventional ventilation during major surgery decreases postoperative pulmonary complications. DESIGN, SETTING, AND PARTICIPANTSSingle-center, assessor-blinded, randomized clinical trial of 1236 patients older than 40 years undergoing major noncardiothoracic, nonintracranial surgery under general anesthesia lasting more than 2 hours in a tertiary hospital in Melbourne, Australia, from February 2015 to February 2019. The last date of follow-up was February 17, 2019.INTERVENTIONS Patients were randomized to receive a tidal volume of 6 mL/kg predicted body weight (n = 614; low tidal volume group) or a tidal volume of 10 mL/kg predicted body weight (n = 592; conventional tidal volume group). All patients received positive end-expiratory pressure (PEEP) at 5 cm H 2 O. MAIN OUTCOMES AND MEASURESThe primary outcome was a composite of postoperative pulmonary complications within the first 7 postoperative days, including pneumonia, bronchospasm, atelectasis, pulmonary congestion, respiratory failure, pleural effusion, pneumothorax, or unplanned requirement for postoperative invasive or noninvasive ventilation. Secondary outcomes were postoperative pulmonary complications including development of pulmonary embolism, acute respiratory distress syndrome, systemic inflammatory response syndrome, sepsis, acute kidney injury, wound infection (superficial and deep), rate of intraoperative need for vasopressor, incidence of unplanned intensive care unit admission, rate of need for rapid response team call, intensive care unit length of stay, hospital length of stay, and in-hospital mortality.RESULTS Among 1236 patients who were randomized, 1206 (98.9%) completed the trial (mean age, 63.5 years; 494 [40.9%] women; 681 [56.4%] undergoing abdominal surgery). The primary outcome occurred in 231 of 608 patients (38%) in the low tidal volume group compared with 232 of 590 patients (39%) in the conventional tidal volume group (difference, −1.3% [95% CI, −6.8% to 4.2%]; risk ratio, 0.97 [95% CI, 0.84-1.11]; P = .64). There were no significant differences in any of the secondary outcomes.CONCLUSIONS AND RELEVANCE Among adult patients undergoing major surgery, intraoperative ventilation with low tidal volume compared with conventional tidal volume, with PEEP applied equally between groups, did not significantly reduce pulmonary complications within the first 7 postoperative days.

Abstract. This paper demonstrates the impact of integrating session types and object-oriented programming, through their implementation in Java. Session types provide high-level abstraction for structuring a series of interactions in a simple and concise syntax, and ensure type-safe communications between distributed peers. We present the first full implementation of a language and runtime for session-based distributed programming featuring asynchronous message passing, delegation and session subtyping, combined with class downloading and exceptions. The compilation and runtime framework of our language effectively maps session abstraction onto underlying transports and guarantees communication safety through static and dynamic session type checking. We have implemented two alternative protocols for performing correct and efficient session delegation and prove their correctness. Benchmark results show the session abstraction can be realised with minimal runtime overhead.

Abstract. Event-driven programming is a major paradigm in concurrent and communication-based programming, and a widely adopted approach to building scalable high-concurrency servers. However, traditional event-driven programs are more difficult to read, write and verify than their multi-threaded counterparts due to low-level APIs and fragmentation of control flow across disjoint event handlers. This paper presents a Java language extension and a novel type discipline for type-safe event-driven session programming that counters the problems of traditional event-based programming with abstractions and safety guarantees based on session types, while retaining the expressiveness and performance characteristics of events. The type discipline extends session types and their primitives with asynchronous input, session typecase and session set types, ensuring eventhandling safety and event progress in addition to the standard type soundness and communication safety. The advantages, expressiveness and performance of event-driven session programming are demonstrated through a range of examples and benchmarks, including a session-typed SMTP server.

Abstract. This paper proposes a new hybrid session verification methodology for applying session types directly to mainstream languages, based on generating protocol-specific endpoint APIs from multiparty session types. The API generation promotes static type checking of the behavioural aspect of the source protocol by mapping the state space of an endpoint in the protocol to a family of channel types in the target language. This is supplemented by very light run-time checks in the generated API that enforce a linear usage discipline on instances of the channel types. The resulting hybrid verification guarantees the absence of protocol violation errors during the execution of the session. We implement our methodology for Java as an extension to the Scribble framework, and use it to specify and implement compliant clients and servers for real-world protocols such as HTTP and SMTP.

Abstract. This work extends asynchronous multiparty session types (MPST) with explicit connection actions to support protocols with optional and dynamic participants. The actions by which endpoints are connected and disconnected are a key element of real-world protocols that is not treated in existing MPST works. In addition, the use cases motivating explicit connections often require a more relaxed form of multiparty choice: these extensions do not satisfy the conservative restrictions used to ensure safety in standard syntactic MPST. Instead, we develop a modelling-based approach to validate MPST safety and progress for these enriched protocols. We present a toolchain implementation, for distributed programming based on our extended MPST in Java, and a core formalism, demonstrating the soundness of our approach. We discuss key implementation issues related to the proposed extensions: a practical treatment of choice subtyping for MPST progress, and multiparty correlation of dynamic binary connections.