Complexity in modern product design is manifest through large numbers of diverse parts, functions, and design disciplines that require an intricate web of synergistic relationships to link them together. It is extremely difficult for designers to assimilate or represent such complex designs in their totality. Few CAD tools provide support for managing complex design relationships. Instead, designers must document and track these relationships independent of the design model. This increases the burden on the design team and increases the risk of inconsistencies and errors in the design. This researches introduces a framework that utilizes the intricate relationships between design components to enhance the representational power of design models and to provide focal points for automating the management of design complexity. Automated mechanisms, based on aggregation, interaction, and variational relationships between design components, are presented that integrate model structure, a v...

Frame Technology was introduced during the early 1980's as an approach to providing significant levels of improvement in software productivity through adaptive reuse. However, despite the claims of order of magnitude improvements and reuse levels of 90%, this technology does not appear to have acquired a corresponding level of popularity, appearing to loose out to the more recent technologies of generative programming and aspect oriented programming. Our interest frame technology stems from our work on the meta-modelling of feature models to support the expression of product line architectures in the UML using a meta-modelling research tool. Whist the tool provides direct support for templates, we sought an approach that would approximate to customisable templates, allowing us to generate specific configurations in response to user choices by severing unused features. Frame Technology appears to offer a potential solution and we present a summary of our investigation of this technology. Finally, we propose future directions of study based around a meta-modelling approach.

Intensional communities are composed of intensional programs sharing a common context, implemented using a networked context server, called the AEPD aether process daemon. The programs are written in the ISE intensional imperative scripting language, which is illustrated through versioned assignments in undergraduate classes. Using this model of intensional communities, two examples are developed: a Web-based class in which a teacher leads a student through a multidimensional page, and the programming of chatrooms.

This paper describes an `external' approach to adding modules to C. The Sloth system consists of a simple scheme for representing C modules as Unix directories, together with shell commands for building and modifying modules and for configuring applications. Sloth modules have procedure definitions, local and exportable declarations, initialization routines and import lists. The Sloth configuration command takes care of all the routine bookkeeping, such as collecting indirectly needed modules, keeping .o files up-to-date (using the standard compiler), generating #includes and ensuring that initializations are performed once each and in an acceptable order. The main weakness is that Sloth is unable to enforce information hiding between modules

The paper describes original research in the area of software configuration management. We assume a software system consists from versions of two kinds: variants and revisions. Our approach is based on defining a suitable model of the software system, formulating requirements for a configuration to be built, and finally building a configuration. The requirements are in fact conditions which constrain the solution. To build a configuration requires search in the model represented as A/O graph. We work with generic configurations which represent solution on a level of variants, and then with bound configurations which represent solution on a level of revisions. We proposed a method to build a configuration which uses heuristic knowledge and implemented it in Prolog. Our experiments show such a method performs better than without supporting knowledge. Keywords: software system configuration, heuristic knowledge, version control 1 Introduction Software system changes more often than usu...

IML is a simple markup language that extends HTML. IML makes it practical for even unskilled authors to produce multiversion Web pages. The form and content of a particular version of a page are determined by the settings of a collection of parameters that specify (for example) layout, in terms of font and size, color, and background color; they can determine sections to be included or omitted, in particular, as droptext; which of a series components to present, and in which order; or they can determine a choice of language. An important advantage of the parametric approach is that members of a community can share pages by sharing parameters.

Abstract. With the increasing internationalization of the World-Wide Web comes a corresponding increase in the need for multi-variant hypertext documents|families of documents with instances that are designed based on the same theme and structure but in which the content varies, being expressed in di erent languages to thereby be accessible in the reader's native languages. The di erent variants in the document family ideally share the same hypertext structure, called a consistent state, but require exceptions to handle variant-unique contents. As a practical matter, di ering maintainer expertise in the supported languages often means that each variant is maintained separately from the others. In this case over time the hypertext structures of the di erent variants gradually deviate from each other, resulting in inconsistencies. We have developed a tool to aid authors in maintaining consistency within multi-variant hypertext documents. Our tool can traverse and compare multiple variants of a hypertext document simultaneously, report inconsistencies, and

What is text? Beyond, that is, material ‘tagged ’ as CDATA in the mark-up syntax jargon. Is it just undifferentiated strings of bytes? Or of ‘characters’?... But then, what is ‘a character’? These are the most basic of many important questions whose investigation is timely in light of the rapid global spread of the ‘XML paradigm ’ for documents and networked applications of all types. They lie behind our more radical questioning of current models of text strings in software, asking whether these are adequate even for ‘alphabetic scripts ’ such as those used in Europe, and ‘syllabic scripts ’ used, for example, in Southern Asia, let alone for the ‘logographic ’ or ‘ideographic scripts ’ of Eastern Asia. Therefore this paper launches a study of ‘text in computers’, aiming first at a basic understanding of useful abstractions for ‘characters ’ and ‘character-strings ’ and asking whether the same abstractions can be usefully applied to all ‘natural languages’? At an implementation level, through this programme we shall determine how best to represent the true nature of ‘text strings ’ in an application-independent way. This is the important first step on the road to developing the many complex models and algorithms that are required to solve current problems posed by the desire to improve the automated handling of all text, in all contexts and in all its aspects. Our resulting interfaces and implementations will certainly make

this paper. A more detailed description is reported in (Bielikov'a, N'avrat, 1995b).

There are two fundamental trends in the development of computers: the miniaturization of components and the increase in communication capacities. The combination of these two trends is leading to a qualitatively new situation, in which the same techniques will be applicable at all scales of computing, be they at the chip level or at the level of the World Wide Web (WWW). 1. Introduction The von Neumann model of computation 3 has been the standard model of computation since the late 1940s. Under this model, a program is stored in a memory and executed by the combined action of a control unit and an arithmetic unit, with loads from the memory and stores to the memory being effected and coordinated by the control unit. This model of computation has shown itself to be remarkably successful. Even today, the vast majority of computers are uniprocessor machines that strongly adhere to the von Neumann model. As a result, the standard paradigm for software is imperative programming, a...