An Experimental Implementation of the Tilde Naming System

The Tilde naming system identifres files in a distributed computing system in a novel way, providing a consistent mechanism for referencing both local and remote files that is independent of the details of the underlying computing environment. The Tilde naming system achieves this independence by providing each user with control of a private name evaluation environment. In contrast, several contemporary distributed systems try to disassociate file names from the structure of the computing environment by hiding complex, heterogeneous computing systems beneath a uniform, global name evaluation environment. The private naming environments supported by Tilde naming incorporate local and remote names into a single naming mechanism while improving software portability and retaining the familiar advantages of hierarchical file naming systems. We have constructed an experimental implementation of the Tilde naming system, and created a computing environment sufficiently rich to support significant software development. With this prototype, we have been able to study the effects of the @ Computing Systems, Vol. 3 'No. 4'Fall 1990 487 Tilde naming scheme on the system's user interface, software project development, information sharing and other issues. This paper summarizes Tilde naming and discusses insights into naming mechanisms gained through the use of the experimental system.