The PODC 2004 conference will be held in St. John's, Newfoundland, Canada, marking the first such event since 1995. Soma Chaudhuri will serve as General Chair, and Krishnamurthy Vidyasankar as Local Arrangements Chair. The 2005 conference's potential European hosting is under discussion. Seth Gilbert and Nancy Lynch discuss Brewer's conjecture that a web service cannot simultaneously guarantee consistency, availability, and partition tolerance. They prove this in the asynchronous model and explore solutions in the partially synchronous model. In the asynchronous model, it's impossible to achieve all three properties. However, any two can be achieved. For example, atomicity and partition tolerance can be achieved with a centralized algorithm. Atomicity and availability can be achieved in non-partitioned environments. Availability and partition tolerance can be achieved without strict consistency, such as in web caches. In the partially synchronous model, messages are either delivered within a known time or lost. While it's still impossible to guarantee atomic consistency in all executions, weaker consistency conditions like t-Connected Consistency can be achieved. This allows for stale data when messages are lost but ensures consistency returns once partitions heal. The t-Connected Consistent algorithm uses a centralized node to manage sequence numbers and broadcasts values, ensuring partial order consistency. The algorithm satisfies all criteria for t-Connected Consistency, providing a practical compromise between consistency and availability in real-world systems.The PODC 2004 conference will be held in St. John's, Newfoundland, Canada, marking the first such event since 1995. Soma Chaudhuri will serve as General Chair, and Krishnamurthy Vidyasankar as Local Arrangements Chair. The 2005 conference's potential European hosting is under discussion. Seth Gilbert and Nancy Lynch discuss Brewer's conjecture that a web service cannot simultaneously guarantee consistency, availability, and partition tolerance. They prove this in the asynchronous model and explore solutions in the partially synchronous model. In the asynchronous model, it's impossible to achieve all three properties. However, any two can be achieved. For example, atomicity and partition tolerance can be achieved with a centralized algorithm. Atomicity and availability can be achieved in non-partitioned environments. Availability and partition tolerance can be achieved without strict consistency, such as in web caches. In the partially synchronous model, messages are either delivered within a known time or lost. While it's still impossible to guarantee atomic consistency in all executions, weaker consistency conditions like t-Connected Consistency can be achieved. This allows for stale data when messages are lost but ensures consistency returns once partitions heal. The t-Connected Consistent algorithm uses a centralized node to manage sequence numbers and broadcasts values, ensuring partial order consistency. The algorithm satisfies all criteria for t-Connected Consistency, providing a practical compromise between consistency and availability in real-world systems.