Collecting research data in the field is still a significant technical problem. While solutions exists, they are all functional rather than elegant. In this project we aim to investigate latency hiding as an elegant underlying formalism for field data collection.
Latency hiding is the idea that a computer program acts as if it is directly connected to a central server but is able to work even when the latency to that server (i.e. the time it takes data to get to that server) is infinite. The underlying latency of the system is hidden from the program user and also hidden from the program developer. Key in this concept is that "offline" no longer means "can't connect to server", it instead means "unknown (and probably very long) latency".
Develop a formal semantics for latency hiding connecting a field data recorder to a central document store. By creating this semantics we hope to validate existing implementations of latency hiding or to highlight potential problems. The formal semantics will also allow latency hiding to be implemented in new systems. We also hope to illuminate the consequences of different design decisions such as what approach is used to resolve data races.
We ran a pilot for field data collection in S2 2016, creating tooth-tan as a proof of concept. Tooth-tan will constitute the basis of our formal modelling work.
There are field data collecting apps used all over the University. We will collect ones we know about here because at the end of the project we will relate our work to each of these:
- Children's Hospital Incontinence Web Site