It is difficult to build user interfaces that must be distributed over a set of dynamic and heterogeneous I/O devices. This difficulty increases when we want to split, merge, replicate, and relocate the UI across a set of heterogeneous devices, without the application intervention. Furthermore, using generic tools, e.g. to search for UI compoments or to save/restore them, is usually not feasible. We follow a novel approach for building UIs that overcomes these problems: Using distributed file systems that export widgets to applications. In this paper we describe Omero, a UI server built along this line for the Plan B Operating System.

Commercial UNIX window systems are unsatisfactory. Because they are cumbersome and complicated, they are unsuitable companions for an operating system that is appreciated for its technical elegance. Their clumsy user interfaces clutter the view of the operating system. A good interface should clarify the view, not obscure it. Mux is one window system that is popular and therefore worth studying as an example of good design. (It is not commercially important because it runs only on obsolete hardware.) This paper uses mux as a case study to illustrate some principles that can help keep a user interface simple, comfortable, and unobtrusive. When designing their products, the purveyors of commercial window systems should keep these principles in mind. Introduction Mux is a window system with no icons, no help facility, no customizability, no noise, and only two menus (one with five entries, one with seven). The spareness of its user interface distinguishes it from commercial window system...

A graph browsing system has been extended to support several general graph operations related to topological graph theory, such as Cartesian product, suspension, face tracing, genus distribution, crosscap distribution, and converting between graphical rotation projections and combinatorial rotation systems. Facilities have also been provided for generating topological building blocks such as bouquets, cubes, paths, cycles, and complete graphs. This system can be used to define graphs diagrammatically and explore their topological properties. The graph browser, called Bonsai, provides an interpreter for a simple graph description language and a graphical interface for modeling information in the form of graph and set diagrams. Graph structures can be created with the browser and converted into a variety of textual forms to interface with other applications. Similarly, separate applications can generate text streams and direct them into Bonsai's interpreter to display the resulting anima...

The Iris library provides a basis for constructing user interfaces on bitmap displays. Written in C++, it provides a window class as the main data type, along with support for such graphical classes as point, rectangle, font, icon, etc. Windows can be created within windows; this property, along with the derived class notion in C++, allows the user to design complex interface objects built from simpler components in a modular fashion. In addition, a window can specify which input events, ranging from a change in a specific mouse button to a move request, it will accept. This paper provides an introduction to Iris and a discussion of its use.

Abstract not found

This paper describes the low-level primitives necessary to implement a particular flavor or inter-process communication. It is motivated by the design of a communication subsystem for a higher-order functional language [Cardelli 84]. Although we try to abstract somewhat from the special characteristics of that language, the model does not accommodate a wide range of communication schemes. This communication model is intended to be used on (uniprocessor) personal computers. In this model, processes running on the same processor can share the same address space. If the underlying language is safe, a process can affect other processes only by communication, or by affecting data structures which have been explicitly transmitted. This ensures privacy and data protection even in a shared address space. Processes running in the same address space can exchange arbitrarily complex objects very cheaply, just by passing pointers. Processes running on different processors communicate through restricted "flat" channels, e.g. character channels. In this case, complex objects have to be encoded to fit into flat channels, and decoded on the other side; the encoding activity may or may not be automatic. In any case there is a semantic difference between exchange of objects in the same address space, where objects are shared, or in different address spaces, where objects are copied. The basic communication mechanism is rendezvous [Milner 80]: both the sender and the receiver may have to wait until the other side is ready to exchange a message. Both the sender and the receiver may offer

this memory behavior for extensive low-level traces collected on a DECstation 3100 workstation. This discussion motivates the development of different caching strategies studied in Chapter 7. Finally, Chapter 8 summarizes the contributions of this research.

This manuscript presents a survey of the theoretical foundations of regular expres- sions and text editors. A thourough introduction to the subjects of symbols, lan- guages, grammars, regular expressions, finite automata, character sets (Unicode), pattern matching, and the inner and outer workings of a text editor is given. The knowledge contained in this survey is then put to good use in presenting a method of compiling regular expressions into efficient finite automata that are well suited for pattern matching tasks in a text editor. The usefulness of pattern matching inside a text editor is also shown, by describing the command sets of modern text editors, such as sam and vim. Finally, a pattern-matching library has been written that can be used in, for example, a text editor. A library implementing an efficient data structure for rep- resenting sequences has also been written. Both these libraries have been used to create an example text editor that uses the pattern-matching library in implement- ing its command set and the data-structure library in implementing its buffering strategy. All pieces of software have been documented using a literate programming style of writing and are included in this manuscript for reference.

M. Ritchie In a new version of the Unix operating system, a flexible coroutine-based design replaces the traditional rigid connection between processes and terminals or networks. Processing modules may be inserted dynamically into the stream that connects a user’s program to a device. Programs may also connect directly to programs, providing interprocess communication.

The modern graphics processing unit (GPU) is the result of 40 years of evolution of hardware to accelerate graphics processing operations. It represents the convergence of support for multiple market segments: computer-aided design, medical imaging, digital content creation, document and presentation applications, and entertainment applications. The exceptional performance characteristics of the GPU make it an attractive target for other application domains. We examine some of this evolution, look at the structure of a modern GPU, and discuss how graphics processing exploits this structure and how nongraphical applications can take advantage of this capability. We discuss some of the technical and market issues around broader adoption of this technology.