Adrian Demaid

Trans-organizational innovation: a framework for research

Information and communication technologies, the evolution of a pattern of innovation based on technology fusion and the transition towards a knowledge-based economy are dominant trends. These trends support trans-organitational innovation, which typically involves the design of complex and technologically hybrid products. Trans-organizational innovation involves generating new knowledge out of knowledge inputs which are distributed across disciplines and organizations which may be geographically dispersed. This is critically dependent on management processes associated with learning. Learning is a contextually situated and interaction-intensive process, which during product innovation involves mutual interaction between characteristics of product and those of context. Such interactions continually evolve the designed form and functionalig of a product. Existing research has tended to neglect the complexities involved in trans-organitational innovation. A product-in-context framework for analyzing the tran...

Prototype-oriented representation of engineering design knowledge

In this paper we present arguments concerning the nature of inheritance hierarchies for representing the relationships between computing objects which encapsulate engineering design knowledge. We review and criticize some other options among object-oriented languages and knowledge representation systems for creating inheritance hierarchies. These are languages and systems which support the relatively rigid class/instance model of property sharing and use a support environment for modification and change. By contrast we provide coupling between objects which employs a scheme of property inheritance that uses ‘exemplary’ or prototypical objects. Our approach yields a class-free inheritance system, in the tradition presented in the literature or the Actor languages, that enables design concepts to emerge easily as the user describes and modifies them at the keyboard. Our prototype-oriented approach offers a means of continually replicating design concepts through object refinement: this principle of one object being specified as a refinement of another object is specialization . We show that treating a knowledge base as evolutionary encourages exploratory comparisons and supports its customization for design representation. We use single taxonomic inheritance to demonstrate the effectiveness of representing information which originates from different perspectives (distinguished structurally as separate inheritance subhierarchies which allow a design world of discourse consisting of many ad hoc groupings to evolve) and criticize the use of multiple inheritance for knowledge representation.

Software machine designed for selection

A knowledge description language designed to be useful in tackling the automation of real-world selection decisions is introduced. Such selection decisions typify the conceptual stage of engineering design, and are exemplified in the paper by reference to the problem of selecting polymers for manufacturing containers. The language is used to create a classification hierarchy, which may be construed as a collection of objects related in terms of a graph which determines a pattern of default property inheritance. The language provides for the formulation of symbolic descriptions of an objects and its properties, and allows the graph to evolve in a controlled interaction with the computer user. Each fresh description of an object is treated as a partial or incomplete description, which is subjected to a strategy for maintaining the consistency of all objects in the graph. The result is an apparatus in which objects inherit dynamically, are always mutable, and which abstract both state and procedural properties.

Classifying Design and Design Management in Seasonal Industries

This paper classifies design processes in seasonal industries, where designs of consumer products have to be delivered to tight deadlines, from the viewpoint of the relationship between the design company, the retailer and the final customers. The paper looks specifically at the risk that designers and retailers carry, the design rework required by the retailer and the communication between the retailer and the designer. We employ a motoring metaphor to show six models of design.

Object-oriented representation of qualitative engineering properties

Abstract The paper presents objections to the prevailing record-oriented data modelling of physical-world properties of use to engineering design. The limitations, which produce an oversimplification in the representation of property information, are highlighted in relation to software systems for materials selection. The overarching problem lies in the highly context-specific way in which properties of materials and engineering products need to be interpreted. By way of redress, two techniques are presented which are supported by a special object-oriented language devised to process design descriptions. The first technique provides a representation of properties of families of materials as distributions of numerical values within ranges which are automatically generated by family members. The second technique involves treating qualitatively described properties as agglomerations of property features compounded or together. Both techniques are supported by a special language specification of complex prototypical objects which depict the properties of any given entity themselves as entities.

Message passing between individual and socially acquainted objects in Smalltalk

This paper reports additions to the commercial, object-oriented language, Smalltalk-80 and their incorporation into a knowledge based environment, POISE. These additions are suitable for the purposes of knowledge representation of engineering design. A dynamic, knowledge representation scheme is supported that allows temporary associations between objects residing in separate, static, hierarchical structures. Message passing between these associated objects is dynamic. Messages are passed, incompletely satisfied, between socially acquainted objects in order to complete a computation. We show that a computing paradigm previously achieved by a specific language, written in the Actor tradition, can also be achieved in a strictly class-instance based language which is then used to create a design knowledge representation environment. A simple, accessible example is used to illustrate the power and generality of the new language.

Classification of plastics materials

Abstract In this paper we report the development of a class-based, hierarchical inheritance structure designed to classify plastics materials. The classification is intended to define and order those families which summarize physical similarities between grades of polymers which are relevant to the engineering designer by virtue of chemical similarity. Our intention in creating this principled classification is to design a structure for an engineering plastics representation which is amenable to extension and mutation while at the same time resulting in useful and practical groupings of the database members, the plastics grades themselves, from the point of view of an engineering designer seeking materials information We describe several alternative approaches to the problem of factoring plastics materials into classes as dictated by the domains characteristics and demonstrate that while a classification based on criteria of chemical structure is a prerequisite for an extensible data-structure, no such classification will factor these materials in such a way that they consistently group like with like from an engineering applications viewpoint. This is explained by the fine control over properties which results from minor chemical changes or additions exercised by the design chemist, and microstructural and geometrical changes arising during processing. Although our example is plastics materials, the tailoring of specific properties for a particular, useful end is common to all materials representations. We criticize classifications used in design representation software and existing software mechanisms in the object-oriented domain. A solution to the problem of maintaining a general chemical classification and of factoring plastics information for engineering use results from a detailed analysis of the property structures used to describe a plastic. A new software mechanism which introduces properties relevant to specific applications at a grade, or instance, level without offending the inheritance hierarchy used to convey generality is presented.

Access Enhancement Objects for data management in Smalltalk

This paper reports an approach to persistence in object-oriented languages, such as Smalltalk, that require a memory-resident image for computation. The architecture uses a selective persistence that provides sufficient data-handling, with support to evolutionary data description and data consultation, for an engineering domain application without any special syntactic extension to the application language. The approach is oriented to knowledge representation and has been tested on materials data information, from several sources, that informs engineering design modelling. The architecture described is an extension of Smalltalk that uses Access Enhancement Objects, which we term Persistence Enhancers, for data management, in conjunction with a storage model optimized for domain modelling.

Reverse knowledge engineering: a methodology for knowledge management

In this paper we discuss an experiment to recreate the evolution of a design in order to elicit the progression of knowledge during the course of a project. The project being studied featured an innovation-the creation of a combined telecommunications and sensing network by a multi-national company in cooperation with other large and small enterprises. The project took five years to complete, during which its scope expanded, the people and companies involved changed, and their numbers increased. The key technologies of the project involved purpose specific hardware and software, all of which needed to interact with an existing telecommunications network. We took it as axiomatic that knowledge was created during the course of the project and that our difficulty was to avoid post hoc rationalisation of the processes involved. To solve this problem we created an exercise in reverse knowledge engineering (RKE) focussed on a prototype-a box of electronics hardware containing embedded software and the capability of interacting intelligently with a network. A panel of engineers and researchers traced the evolution of knowledge during design changes by reverse engineering the product, essentially dismantling it and using the parts as a touchstone for building a picture of knowledge evolution. We report the benefits and pitfalls of the procedure, drawing particular attention to its ability to identify the trace elements of redundant decisions, these being the most difficult to identify by conventional interview techniques.

The Design of a Networked Product by a Network of Product Designers the Design and Implementation of Services Delivered Across a Network is a New Art

Products and services containing hardware and software that are delivered across a network are expensive to design and carry a particularly high risk when new markets are being created in a highly competitive, global marketplace. The design and implementation of such systems in a competitive industrial environment is a relatively new art that requires distinctive skills. The rate of change in the market is so high that the resulting time pressures are changing design-to-market processes in traditional industries. Our interest is in Telecommunication Network-based Services (TNSs), a term used to describe a range of different services that rely on additions to a general communication network, such as the public telephone system, to deliver added value in the form of new products. Resulting from in-depth studies we report the design of a remote meter-reading system across a network of competing and collaborating companies. We comment on the way in which design-to-market processes are changing in our collabor...