Engineering ontologies

Engineering ontologies

In any field, the choice of words used is important in efficient communication. When we speak or write, we need to be clear about the concepts we have in mind, what we mean when we use particular words, and what the recipient may interpret from the communication. In the high-tech, high-precision world of engineering, clear precise communication between engineers is every bit as important as the tolerances of the components going into the car, aircraft, and medical equipment they may be building.

Shorter lead times, increasing sophistication of products, and the complexity introduced by wider geographical distribution of engineers working on any given project, is only increasing the difficulties in communication. Take for example, the aircraft built by Airbus that are designed by engineers in a number of European countries, or the new X series Jaguar cars designed in the UK, that are based on the Lincoln floor pans designed by Ford in the US, or the range of Rolls Royce aero engines designed in the UK that need to conform with the range of Boeing aircraft designed in the US.

To lubricate this communication in engineering, there is a need for specialist ontologies for defining and describing the concepts and terminology in use in groups, and between groups, of engineers. This need goes well beyond the frameworks provided by standards agencies, such ISO and BSI, and the professional bodies such as the Institution of Electrical Engineers.

A surprisingly simple yet effective approach for helping meet this need for meta-knowledge is the format of taxonomies. These have been used in many fields for a long time. For example, in botany, taxonomies are used for plant classification. Here the complexity may be in the shear number of entities that need to be considered rather than the need for sophisticated constructs. There are many species of plant, but the ontology is based on a relatively small number of groupings such as phylum, class, family, etc, with "Type-Of" used as a relationship.

Obviously, taxonomies are the basis of classification schemes and indexing systems in information management such as the Dewey Decimal System. Taxonomies are even more wide spread with applications including post codes (zip codes) used by postal services, and job categories used by tax collection agencies. With the advent of the internet, there has been increased interest in using taxonomies for structuring information for easier management and retrieval.

Consider a navigation path in an online catalogue that starts at computers, goes to multimedia computers, and ends at multimedia computers at home. Now consider another path starting at home electronics, goes to home entertainment, and ends at home multimedia computers. Probably, these two end points refer to the same item. By adopting a common ontology, the use of different terms for the same thing may be obviated or acknowledged as required.

In engineering companies, parts catalogues can be a key component of an engineering ontology. Obviously, engineering companies have been using parts catalogues for a long time. However, there is now the opportunity to do more with them in terms of knowledge management and operations management. This is leading to the need for better developed classification schemes that can work in different organizations. A number of "standardized" classification systems have been proposed to allow companies to work together, such as such as by the Electronic Commerce Code Management Association (, though at the moment there are no dominant proposals.

However, Findlay Publications (, a specialist publisher of engineering magazines and directories, has made significant progress in developing an online classification scheme that works across engineering sectors, and across suppliers. Originally, this classification scheme was built in-house for use in organizing its directories, in particular the Design Sector Global (, a design engineers selection and specification tool containing product details from thousands of of manufacturing suppliers.

At Findlay Publications, a team of engineers, technical editors and professional classification experts spent 18 months examining all available schema noting the best features of each. In addition, software tools were developed to support both the editing and the use of the classification system. The resulting classification scheme incorporates over 17,000 product categories which are broken down to a current maximum of 11 levels. Each category may be further categorized using attributes by drawing on a selection of the 1700 attribute types incorporated in the system.

Given the potential benefits of an engineering classification scheme (See Box 1), Findlay Publications are now looking to collaborate with other organizations in adapting their classification scheme for other users. The combination of their engineering and editorial expertise puts them in a good position. Already, it is being applied by the UK Society of Motor Manufacturers and it is the basis of a major US online venture in the automotive sector.

A number of software supplier have developed products to support the creation and management of taxonomic catalogues for B2B ecommerce. A leading specialist supplier of catalogue content management software is Requisite Technology ( based in Colorado in the US. Products include eMerge which allows organizations to construct an online catalogue for procurement. Suppliers load product information into eMerge, and this information is organized into a consistent structure and staged for review and approval before loading into an eProcurement catalogue. Searching an online catalogue is then via text searching, including key word searching, or tree searching. In tree searching, the user navigates through the taxonomy of categories of items. Organizations using this software include Xerox and various net marketplaces including SciQuest,, and Petrocosm.

Another supplier of catalogue management tools is Intermat ( with their Standard Modifier Dictionary, a widely used standard for the description of maintenance, repair and operating inventories, with customers including Rockwell Automation. The dictionary is based around the idea of using nouns (eg. valve) with modifiers (eg. butterfly) to categorize entries, and to provide definitions, guidelines, characteristics, and synonyms, to help users navigate the dictionary. Currently they support over 2,400 noun-modifier pairs. Further examples of classification systems for comparison can be found at,, and

Within knowledge management, the role of taxonomies can be pushed wider to give richer engineering ontologies. An example of this is the technology taxonomy developed by the UK Ministry of Defence, and used by companies including BAE Systems, to support the development of technology strategy. The aim of the technology taxonomy is to allow technology developments to be related with potential applications and eventually with military equipment. So advances in technologies (eg. IR absorbing material or radar absorbing material) are linked with primary applications (eg. stealth designs) and to resulting equipment (eg. surface ships or fixed wing aircraft). The taxonomy incorporates over 90 technology areas and over 80 application areas.

When looking at examples of taxonomies, one of the things to note is the frequent use of relationships such as "Type-Of" and "Part-Of". This gives an important handle on ways of constructing taxonomies. Whilst, there are no hard and fast rules for constructing taxonomies, in a number of disciplines, including artificial intelligence and computational linguistics, there are a number of proposals for structuring lexical and ontological information. Of particular relevance are meronymic relations (Part-Of) relations. These are different to Class-Member relations (eg The New Forest is a forest), and include Component-IntegralObject relations (eg handle is part of a cup, Portion-Mass relations (eg a slice is part of a pie, Stuff-Object relations (eg gin is part of a martini, and Feature-Activity relations (eg paying is part of shopping. Harnessing a richer breakdown of meronymic relations can create interesting opportunities in building engineering ontologies.

Of course, the development of an engineering ontology does not need to start from scratch. At Boeing (, a considerable investment has been made by Boeing Technical Libraries in developing by hand a technical thesaurus in the form of a semantic network. This incorporates 37,000 concepts with an additional 19,000 synonym concept names, and 100,000 links including broaderTerm, narrowerTerm, and relatedTerm. Examples of entries are jet engines broaderTerm engines and engines relatedTo propulsion systems. Most of the concepts are restricted to Boeing and aerospace, and the semantic network was intended to be used by humans.

Recently, a requirement for an expert locator system arose in Boeing. The company has a large workforce of experts but knowing who is an expert in what can be difficult. Therefore, this requirement was for a web-based system which can be used by typing in the type of expert being sought, and the system would return details on potentially appropriate experts.

In order to develop an effective solution, the Boeing technical thesaurus was harnessed as the knowledgebase for the web application. However, there is a considerable difference between a human using a thesaurus and a computer using it. Humans bring commonsense and general knowledge to fill gaps in the thesaurus and for interpreting the information in it. To address these short comings, the developers harnessed some basic techniques form artificial intelligence to refine the thesaurus in particular with regard to missing links in the semantic network to give a useful online engineering ontology that contains many complex concepts at a fraction of the cost of developing one from scratch.

Motivations for engineering ontologies are diverse. Key amongst them are the need for: Supporting more reliable, less ambiguous, communication; Improving searching through large amounts of structured or unstructured text; Supporting automatic classification of text; Providing automatic indexing; and Facilitating the creation of linkages between related data. Clearly many of the established techniques in information management for thesaurus and classification construction are highly relevant here.

However, the computer-based nature of online engineering ontologies creates both problems and opportunities. Problems come from the need to formalize any assumptions made in the ontologies - since computers are inherently unintelligent, they need everything to be explained in precise terms. Opportunities come from the potential of applying artificial intelligence techniques such as inheritance hierarchies and taxonomies, and for automated algorithms for checking the structure and completeness of ontologies, as successfully illustrated in a range of organizations involved in engineering.

Anthony Hunter is a senior lecturer in computer science at University College London. He can be contacted at:

Box 1: Potential payoffs from a classification system

A comprehensive technical engineering classification system such as the Findlay Classification System is an important knowledge management tool that can help address major issues including:

Communication A classification system acts as a universal translator enabling disparate companies or sites to communicate with a common technical language. Different groups of people may have different names for the same thing. For example, a design engineer may refer to a taper roller bearing for use in a wheel, whereas as a service engineer may refer to it as a wheel bearing. Establishing a common method of classification allows companies to process and analyse data automatically.

Duplication of Work Structuring a classification correctly will mean information will be stored in one location. Using a car as an example: springs are used throughout the vehicle i.e. in the clutch, engine, gearbox, instrumentation panel. Using a conventional bill of materials data storage approach a spring would be stored under all these areas, in a manufacturing company that could result in the same spring stored in multiple places under different part numbers, which means increased stock, more floorspace used, more stock reordering, and so inflated costs. Classification systems store information on form or function not application providing one location and so leading to an intelligent reduction in these costs.

Relational Synergy As a classification groups similar categories together those relationships can be exploited for example when looking for nuts and bolts the user will find it in the fastening and joining area with adhesives, welding, rivets, screws etc. This may prompt the user to consider an alternative, better fastening approach to their problem. Analysing similiar products may also prompt supplier base or product rationalisation.

Rapid Information Retrieval A logical hierarchical tree classification system allows data and information to be navigated rapidly and efficiently. A classification system designed by engineers may prove to be more instinctively navigable by other engineers.