Industrial Hypermedia Design.

Wills GB, Crowder RM, Heath I, Hall W.

Abstract.

The amount of electronic information used in an industrial environment is increasing. How to manage the information is becoming a crucial issue. Part of the solution is to use hypermedia, to allow associations to be made within the information space. While there already exists a number of hypermedia design models and methodologies, none address the issues faced by the industrial designer. This report examines some of the strengths and weaknesses of existing models and then proposes an industrial design methodology that will address the issues faced when designing a large scale industrial hypermedia system.

Technical Report No. M98-2

ISBN:- 0-585432-668-5

Copyright © 1998, University of Southampton. All rights reserved.

Contents.

1. Introduction.
1.1 Microcosm:- Hypermedia in an Industrial Environment.

2. Design of Large-scale Industrial Hypermedia Applications.
2.1 Hypermedia applications.
2.2 Hypermedia Design Model (HDM).
2.3 The Relationship Management Methodology (RMM).
2.4 The Object Oriented Hypermedia Design Model, (OOHDM)
2.5 Hypermedia Design Environments.
2.6 Other Models.
2.7 Industrial Hypermedia Design methodology.
2.7.1 Specification
2.7.2 Structure.
2.7.3 The User Interface.
2.7.4 Access method.
2.7.5 Authoring.
2.7.6 Navigation Considerations.
2.7.7 Revision Control.
2.7.8 Hardware Selection.
2.7.9 Design Reviews.
2.7.10 Prototype.
2.7.11 User trials.

3. Conclusion.

1. Introduction.

In many automated industrial environments, some form of information technology is in use. On the factory floor the use of information technology can range from computer controlled systems with electronic help manuals, to fully integrated electronic management systems. In addition, many companies are intent on implementing a company-wide integrated computer system, with larger organisations using global intranets. The information problems facing industry world-wide are largely the same, the movement of information between design, sales and production facilities

The concept of industrial strength hypermedia was initially discussed by Malcolm [Malcolm 91], since which time the demands on manufacturing enterprises have not abated, especially in areas such as globalisation, virtual enterprises, lean and agile manufacturing. These have made the management of information resources vital to the success of any company irrespective of the product being manufactured.

Traditional paper based information systems have been refined over many years of use. In many cases, an effective document management system exists. As equipment is replaced or modernised, many of the old paper based information systems are being replaced with electronic document management systems.

A successful engineering information management strategy must be able to integrate information resources that are in different formats, and allow personnel with different and varying computer skills effective and controlled access to the required information. For example maintenance personnel are multi-skilled, often working on many different types of equipment/ production lines that are becoming more complex, and breakdown less frequently. Hence, it is not possible for them to remember all the details involved in the maintenance of such equipment. The maintenance personnel therefore have to rely more on procedures, work instructions, manuals, engineering drawings, personnel notes, etc. This highlights a number of problems and limitations with conventional information retrieval techniques [Crowder 95].

• Introduction of errors due to constant cross-referencing between different information systems.
• On older equipment, the quality of factory floor documentation can be poor or incomplete.
• With so much documentation distributed around the factory, there are problems of ensuring documentation are at the correct issue.
• The need to share information gained by experience, by other maintenance teams/personnel is at best very limited.

In addition, the integrity of the information must be maintained by the implementation of suitable change control, and associated security measures.

As the volume of available electronic information increases, managing the information and ensuring that user can access the information quickly and effectively becomes a critical issue.

Crowder et al. [Crowder 96b], have demonstrated that, by using an open hypermedia system, an information resource base can be used for operator training, planned maintenance, and diagnostics. They have also demonstrated how such a system can overcome some of the common problems with traditional computer- or paper-based systems, which present and distribute text and engineering drawings sequentially [Crowder 96c]. This report presents the research into a hypermedia design methodology for managing the growing number of electronic resources in a manufacturing environment.

Back to Contents List.

1.1 Microcosm:- Hypermedia in an Industrial Environment.

Research on open hypermedia has been undertaken at the University of Southampton since 1989 [Fountain 90]. This has result in the development of a commercial package, called Microcosm [Multicosm]. Microcosm is the software package used to create the hypermedia application and will form the basic hypermedia system for this research.

As stated earlier this research proposes a factory wide strategy for managing the growing number of electronic information resources in a manufacturing environment by using hypermedia. A paper by Davis et al. [Davis 92] was produced in direct response to Malcolm, proposing some Microcosm solutions to the problems of using hypermedia generally.

Microcosm consists of a number of autonomous processes, which communicate with each other by a message-passing system [Heath 96]. Microcosm philosophy is that hypermedia links in themselves are a valuable store of knowledge. No information about the links is held in the document data files in the form of mark-up. Instead, all data files (text documents, video clips, digitised photographs, tables, technical drawings, audio files, etc.) remain in their native format, while the link information is held in link databases (linkbases). This enables the documents to be stored in different locations, in their original application format. Microcosm allows the user to navigate through the documentation resource base via a number of different link mechanisms [Davis 94, Heath 94].

The links can be arranged in linkbases to represent different cognitive and pedagogical structures. The open architecture also allows different processes to be incorporated into a hypermedia application. The processes range from simple link creation and search algorithms to dynamic link creation, such as information retrieval and rule-base algorithms [Heath 94].

Back to Contents List.

2 Design of Large-scale Industrial Hypermedia Applications.

The design process can be classed as a creative process involving a multi-discipline team of personnel. The actual design process will involve aspects of authoring and HCI, which are cover fully in other technical reports [Wills 97a, Wills 97b]. The general requirements for a large project management for a large-scale hypermedia application are those of any engineering project. There are of course an even greater number of similarities between large-scale software projects and that of large-scale hypermedia applications. In fact the large-scale hypermedia application can be thought of as a special case of large-scale software project. There are copious amounts of literature on this subject of project management [Chapman 87, Thayer 90], and therefore only those areas that are amplified by the unique character of large-scale industrial hypermedia design will be expounded on in this report.

2.1 Hypermedia applications.

A hypermedia system is one in which the concepts of hypertext are applied to multimedia data. Hypertext is a term coined by Ted Nelson in the late 1960’s [Conklin 87]. It was applied to unstructured text, with associations between the text being made with links. Multimedia can be defined as "Multiple means by which information is stored, transmitted, presented, or perceived" [Hall]. In practice hypermedia allows association to be made between different types of media, i.e. text, video, digitised photographs, databases, engineering drawings, spreadsheets.

While no formal definition of the term hypermedia application exists, the term hypermedia application is widely used to refer to the overall collection of links and documents for a given subject [Goose 96]. For this research the term hypermedia application is used to refer to a collection of multimedia documents, links, and microcosm filters.

2.2 Hypermedia Design Model (HDM).

Garzotto et al, suggest a design model for hypertext [Garzotto 93] called the Hypertext Design Model (HDM). In their paper they explain that the development of a hypertext system would benefit from a structure and systematic development, especially in the case of large scale and complex applications. Their paper presents HDM, ‘a first step towards defining a general purpose model for authoring-in-the-large’. According to Garzotto et al. Authoring-in-the-large is to consider the overall application organisation and behaviour [Garzotto 95]. They suggest that authoring-in-the-large allows a description of the overall classes of the information elements and navigational structures of a complex applications to be obtained without being concerned with implementation details or the system in which it is to be used. Some of the innovative features and terminology used in HDM are:

• Entities: These denote a physical or conceptual object of the domain. These consist of a hierarchy of components, which in turn consists of units. These units are the smallest chunks of information, commonly called nodes in hypertext.
• The notation of perspective; in HDM gives consistency in presenting the information of an entity type in different ways.
• The definition of different categories of links:
  • Structural links: connect together components belonging to the same entity
  • Perspective links: connect together the different units that correspond to the same component.
  • Application links: denote arbitrary, domain dependent relationships and connect together components and entities of the same or different types in arbitrary patterns set-up by the author.
• The ability to define a significant number of links from the conceptual-design level description (especially structural and perspective links that can be deduced from the structure or algorithmically derived).

HDM can be used in different manners:

• As a modelling device: to support producing high level specifications of existing or to-be-developed applications. It is particularly suited to applications where regularity, organisation, modularity and consistency are important.
• As an implementation device: it is the basis for designing tools that directly support application development.

HDM provides a framework for describing and analysing hypermedia applications. HDM is a design model of the hypermedia application and does not explicitly set out the methods required to create the application that is it is not a design methodology [Lowe 98]

Back to Contents List.

2.3 The Relationship Management Methodology (RMM).

This methodology was first proposed by Isakowitz et al. [Isakowitz 95]. RMM is suited to hypermedia applications that have regular structure, those with classes of objects that have a definable relationship, and where there are multiple instances of objects between each class [Lowe 98].

RMM is particularly suited to hypermedia applications where the classes have volatile data that regularly requires updating i.e. product catalogues. RMM is not appropriate for hypermedia applications that are highly structured but remain the same for prolonged periods, or hypermedia applications with dynamic structures and volatile (dynamic) data. The RMM can be shown graphically (see Figure 1).

The core of the design methodology is in the three middle steps S1 to S3.

• Step S1: To represent the information domain of the hypermedia application via an Entity -Relationship Diagram. These entities and their relationship will form the basis of the hypermedia application, this will be implemented as nodes and links.
• Step S2: To determine how the information in the entity is presented to the user, and how the user may access it. This involves splitting the entity into useful slices and grouping them together. The organising of entities into slices is called the slice design phase. The links between the slices are called structural links. Structural links differ from associative relationships in that structural links are between the same entity while associated links link to different entities. The importance of distinguishing between the two is important for navigation.
• Step S3: This step involves designing the navigation paths. To enable updating of the information these are described in generic terms and not hard coded. This is achieved by referring to properties of the entities and their relationships.

Figure 1 RMM Design Methodology

This methodology is mainly concerned with a narrow data set of volatile and structured information. While much of the information in an industrial environment is structured it cannot always be guaranteed to be volatile. Lowe et al. describes the system as "....suitable for applications that have highly structured information models with a lot of information in a collection of related classes (the types of applications which have traditionally been suitable for database applications). These types of applications are often based on a physical system (such as an information system for library loans, or for details on personnel in an organisation)" [Lowe 98]. In addition, they point out that RMM is good for information structure and process development, RMM has not been researched in regards to information re-use, application maintainability, or cognitive management during development.

Back to Contents List.

2.4 The Object Oriented Hypermedia Design Model, (OOHDM)

This methodology was proposed by Schwabe et al. [Schwabe 95, Schwabe 96]. OOHDM builds a hypermedia application using an object-orientated framework. The design method uses a four-step process, each focusing on a particular design control. The four steps are:

• Conceptual (or Domain) Analysis: - The main aim of this step is to capture the domain semantics as naturally as possible, at this stage little concern is made of the type of user or tasks. This results in a conceptual schema built out of subsystem classes and relationships.
• Navigational Design: - The Navigational Design reorganised the information of the conceptual analysis to take into account the intended users who are to perform certain tasks using the hypermedia application. This is achieved by defining a navigational model in terms of nodes, links (reflecting the relationships intended to be explored by the user) and access structures (indexes and guided tours). By taking account of how the user explores the information the author of a hypermedia application is able to avoid the redundant information and prevent the user getting lost in hyperspace. By defining the navigational semantics in terms of nodes and links a model movement in the navigational space can be made independently of the conceptual model.
• Abstract Interface Design: - In this stage perceptible interface objects are defined in terms of interface classes. Interface behaviour is declared by specifying how to handle external and user generated events, and how communication takes place between navigational and interface objects. A clear separation between navigational and abstract interface design allows different interfaces for the same navigational model, giving a higher degree of independence.
• Implementation: - In this final stage the designer maps the navigational and interface models into a concrete object.

The advantage of this model is that the designer can clearly focus on the importance issues in each step. However OOHDM does not cover the early aspects of the Design Process nor does it address how the steps are co-ordinated [Lowe 98]. What is clear from the papers is that an author will have to be totally conversant with Object-Oriented design methods and software. This is not practical in an Industrial environment where the author would be an expert in another field (i.e. Production Engineering, Material specialist, Procurement, etc.).

Back to Contents List.

2.5 Hypermedia Design Environments.

Nanard & Nanard [Nanard 95] have set out the requirements for a design environment. They also state that the range of hypermedia applications is so broad that no single formal design technique is relevant to all of them. They also say that any design environment should support:

• A fast experimental feedback loop, which facilitates the evaluation activity and process recursion, thereby improving design quality.
• Tools for generating and instantiating models, which enable designers to alternate among bottom-up and top-down approach, thus promoting both structure and updating activities.
• Easy and unconstrained cloning tools, which work at the instant level, thus supporting the generating material activity.
• The ability to easily build solutions that may be evaluated directly regardless of the user’s current stage within the formal design technique.
• Integrating light prototyping to evaluate design choices-at least the presentation and the navigation ones- at any stage of development.

They concluded by saying that ‘Formal design techniques improve consistency, but due to the human factors in the design experimental feedback is vital’.

Johnson [Johnson 95] also noted that ‘Regardless of a hypertext’s size, the best quality-control tool is a skilled author. While an automated toolset could generate a comprehensive analysis, it takes an author to distinguish actual errors from design choices’.

Back to Contents List.

2.6 Other Models.

Balasubramanian et al. also rejected the RMM, HDM and OOHDM models when designing a large-scale web site for a financial-management company [Balasubramanian 97]. The reason given was that ' All of them require the application domain to be abstracted in the forms of entities or classes and relationships' This model could not be used in this research as it was design specifically for the application and the Wold-Wide-Web domain.

The design model suggested by Ginige and Lowe [Ginige 97] gives a clear distinction between the design processes and the authoring processes. Ginige and Lowe collected the design processes into a group that they call the Creative Process. This is a multi-discipline stage that will involve Application Specialists, Computer Scientists, Graphic designers, HCI specialists; etc. This overcomes some of the limitations of the RMM by including the specification in the process. They recommend that the specification includes the following inputs:

• Technical issues;
• Content Outline;
• Cognitive issues;
• Non Technical Issues

The output from the design stage should address the following issues:

• Information Structure;
• Information Access Method;
• Look & Feel;
• Guidelines for Content Development.

Figure 2 The Design model suggested by Ginige and Lowe for designing large scale hypermedia applications.

Back to Contents List.

2.7 Industrial Hypermedia Design methodology.

Due to the size and complexity of industrial hypermedia, there are three main areas unique to large scale hypermedia design, in which the complexity of the task and hence the potential for errors to occur are:

• The increased cognitive burden to the author.
• The maintainability of the system. There is now a requirement to maintain the links and ensure that they are still valid if a document is moved or deleted.
• The reuse of information now must take into account the perspective implied by the links.

This research addresses these issues in the context of proposing a design methodology for an industrial hypermedia application. The design methodology needs to be flexible enough to accommodate the current and any future philosophies used by the manufacturing industry. This includes the quality and company controls on the information, management strategies (Total Quality Management, Agile Manufacture, etc.) and the quality requirement on the design procedure.

The information resource management system that is developed must allow the integration and dissemination of information to the engineers, managers and operators working in an industrial environment. In most industrial hypermedia applications, it can safely be assumed that some form of propriety software will be used. In this research as with most industrial applications the propriety software used for the hypermedia application determines the linking mechanism. The use of propriety software can limit the application designers to the paradigms set by the designers of the software. However, it is still possible to be creative in the use of the linking mechanism.

The design of an industrial hypermedia application is an iterative process involving several stages, which are described below, while a summary of the design process is represented by figure 3.

Figure 3 Design Methodology for an Industrial Environment.

Back to Contents List.

2.7.1 Specification

As with any engineering project one of the first tasks is to establish the project requirements and to write a specification. A user specification has been written by the authors for the FIRM project [Crowder 96a]. The specification serves, as a good example of what is generally required for an industrial hypermedia system. The specification was written and published well before the design model suggested by Ginige and Lowe [Ginige 97]. However, all of the broad subject areas recommend by Ginige and Lowe for a large hypermedia application are included. In addition, the authors have added additional requirements specific to the manufacturing industry and in particular the maintenance environment.

2.7.2 Structure.

An industrial environment will generally already have a large established information space, (Figure 3). The design methodology proposed does not replace these existing information systems, but incorporates them. This will ensure that all the essential information that already exists is included into the hypermedia application. In an industrial environment the structure of the information is generally hierarchical and will often relate to physical objects and not just concepts. Hence, the same piece of equipment can be used in different parts of the organisation. In addition, different people, from different departments carrying out different tasks use the same information. In most industrial environments the structure of the information is defined by:

• The structure of the organisation. An organisational chart in the quality manual normally illustrates the organisational structure.
• The job function of the person using the information.
• The layout and use of equipment, process lines, manufacturing cells, etc.

This naturally creates different views of the information depending on who is accessing the information, which task they are undertaking, and what piece (if any) of equipment is being used or worked on. Hence, there is no single structure to represent the information. Using the principles of linking in hypermedia, any number of information structures are possible, reflecting the different views on the information depending on who and what task is being undertaken. Therefore, in the information space shown in figure 4, there are few lines showing the relationship between the islands of information.

However, by using the ‘natural’ hierarchies that exist within the organisation several charts can be produced. These charts can then be used to aid authoring and give an initial structure to the information. These ‘natural’ hierarchies, can be used to allocated responsibility for maintaining the different islands of information. In practice the information will be ‘owned’ by a particular department or job function, for example the Technical Instructions are controlled and written by the engineering department for use in many other departments (i.e. production, maintenance, training, etc). How to structure and link these islands of information is explained below under authoring.

Figure -4 The information space for a process line at Pirelli Cables

Back to Contents List.

2.7.3 The User Interface.

The user interface is the means by which the user will initially judge the application. Hence, the user interface is an essential element in influencing the success of the hypermedia application. It is via the user interface that the designers provide the tools by which the user accesses and navigates the information space, maintain the system and author further hypermedia applications. While the interface may be customised, the designers are limited to the interface available. However Page et al. [Page 96] recommend that expert users be allowed to customise the interfaces, to improve efficiency. In this design methodology, the settings for the customised interfaces will be part of the design output, based on the results of the user trials. Navigation and user interface are part of the subject of Human Computer Interaction (HCI), Wills et al. [Wills 97a] examined some of the issues involved with user interfaces for an industrial hypermedia system.

2.7.3.4 Access method.

This stage of the design examines how the user is to be allowed to access the information space. This is again treated separately although it is closely related to the navigation paradigm. The access to the information space can be controlled by using a series of permissions and resulting menus and/or different linkbases. A user is shown only those options allowed by the access permission given. This can be achieved by using either different menus or different linkbase giving alternative routes through the information space. For example from the drawing of the process line, the operator will not have any links to the maintenance information, while the maintenance personnel will be able to follow links created for both the maintainers and operators.

Back to Contents List.

2.7.5 Authoring.

How to structure and author one of these islands of information is explained in another technical report [Wills 97b]. In this report only details of the design aspect are presented. The largest cost in producing a hypermedia application is authoring. This high cost results from the time spent by a person experienced in the subject area to manually linking the information into a cognitive and pedagogical structure that is easy to navigate [Crowder 97]. Therefore much of the process of authoring needs to be automated, thereby reducing the time required of the ‘expert’ to link the hypermedia application and hence reducing the cost. The hierarchical structure of many technical manuals and procedures comes to the aid of the industrial author as most chapters, sections, and subsections, etc. are formatted using heading styles (or at least are numbered or in bold if not underlined and normally of a different size font). Yet, the majority of authoring effort is spent on producing these structural and explicit links, in an almost administrative role, to link -

• The table of contents (or index) to the relevant section.
• A list of tables (or figures) to the relevant table (or figure) and their reference in the text.
• Explicit references to other procedures, data sheets, etc., be it on the local system, or on the WWW.

Therefore, the majority of these links can be automatic generated using macro languages available in many modern word-processing packages. Automatic generation of links is made easier by the use of templates, procedures or guidelines for construction of the electronic documents. In practice, several templates, procedures and guidelines are required (i.e. for manuals, memos, reports, specifications, etc). Some of these may already exist as part of the company’s quality procedures as they generally improve efficiency and consistency when producing documents. The designers role here is to produce a guideline on how to write the templates and macros to fit into the authoring and design rational.

The information relating to the information on a process line has the advantage that it essentially can be associated with a physical object and not just a concept. The information for these individual components can be viewed as islands of information (see Figure 5). The information for each pieces of equipment can be authored separately in a hypermedia application and consisting of all the necessary:

• Microcosm specific: Filters, Linkbases, and Computed Links.
• Engineering Specific: Manufacturing Data, Documents, Pictures, Videos, Procedure, Safety instructions etc.

As the size of these application are small compared to the eventual size of an industrial application we have called these Mini-hypermedia applications (MHAs). MHAs will save time (and hence cost) in not having to re-author the same information for units at the same revision level and modification state. MHAs also enabled portability and modularization of components and subsystems on the process line. Hence, a Mini-hypermedia may contain several other separately authored Mini-hypermedia applications.

Figure 5 The Diagram shows how the MHA of a Drive can be assembled using pre-authored components from different Manufacturers.

This process is then extrapolated, to include all the other islands of information within the factory information space as illustrated in Figure 4. Therefore, a larger hypermedia application can also be constructed out of a series of pre-authored MHAs. These mini-hypermedia applications can be created ‘off-line’. Hence, the required effort and therefore cost of subsequent authoring is reduced as these MHAs are produced.

Figure 5 shows the structure through the information space to appear hierarchical. However, in practice the growth of the information space will be dendritic in nature.

By creating the larger applications out of mini-hypermedia application (MHA), the author is able to use current tools and techniques normally limited to smaller hypermedia applications, for example link maps. Also the smaller application becomes more manageable and therefore reduces the cognitive burden to the author.

Figure 6 Diagram showing an example of how different linkbses are used to aid maintenability, resuability and authorability of a MHA

Each concept or pedagogical structure is represented by a different group of links, each held in a separate linkbase (figure 6). The links referring to information outside (external to) the MHA are kept separate from those linking information only within (internal to) the MHA. This separation of the links this will also increases the maintainability of the MHA.

In addition the separation of internal and external links enables the danger pointed out by Garzotto et al. [Garzotto 96], when reusing information to be avoided or at least reduced. Garzotto et al. explained that when reusing information containing button links, the meaning of these links could change depending on the context in which they are re-used. For example the reuse of a node in a MHA with a next button (the next item can be a different item in the re-used version).

2.7.5.1 Administrative Overlay.

In order to adopt the MHA approach to authoring a large-scale industrial hypermedia application it was necessary to build an administrative overlay. This allowed the MHAs to be merged into one application, something the current release of Microcosm cannot do. This used the Microcosm system at its core, yet enable extra linking mechanisms and utilities specific to an industrial application. The key features of the administration overlay was that it enabled:

• Access control to meet the requirements for a documentation control system.
• The agglomeration of MHA into one application.
• Control of the importing new and updated MHAs.

In addition, this approach has the advantage that it separates the information from the delivery platform. Therefore, features can be added to the environment without having to change the underlying hypermedia system. In addition, the output is a translator, hence the output can support other publishing systems, for example a set of web documents.

Back to Contents List.

2.7.6 Navigation Considerations.

This is a top-down approach to the design, on how the user is to navigate (move through and around) the information space. Navigation can be viewed as a combination of hypermedia linking, information retrieval and document management.

While Navigation, user interface, and linking mechanism design are closely associated, they are initially treated as separate issues. This ensures that the designers are not limiting their options too soon. It is important that the designer examines ways of navigating the information space that is not dependent on or limited by the user interface or linking mechanism, at least in the initial iteration. This allows the designer to explore and use the natural structures that exist, the context in which the application is to used, the specific user requirements, and paradigms for the use with electronic information. The method of hierarchical structure, described above, allows the user to navigate quickly, providing the levels are not too deep. This method of hierarchical authoring will enable ease of navigation as the links will apply only to the hypermedia application for which it is originally built (child) and any the author imports them into (parent). For example a mould machine manual will have a section on changing the bearings on the mould machine drive unit. This is then imported as part of a MHA for the whole line. However, when the user is working on the mould machine and wanting to change the bearings on a drive unit, they do not want to know about changing the bearings on any other machine (for example the caterpillar drive). Yet, if the user is looking at the line level and queries the system on changing bearings they may want the information on changing the bearing on all the pieces of equipment.

2.7.6.1 Link Clusters

The MHAs related to tangible information resources, and hence have physical structure. In addition, the MHAs can also be linked together using the linking mechanism available in Microcosm to represent pedagogical or conceptual themes within the MHA. However, there needed to be a higher level of linking that would link abstract concepts between the MHAs. This led to the notion of Link Clusters.

Link cluster expands the idea of using Generic links in Microcosm, where Generic links are used to link all of the relevant information about an item together. For example, each document that contained information about the component Y145 say, the author would define a generic link where the source anchor is the name of the item (e.g. Y145), and the destination anchor the location of the name as it appeared in the documentation. This allowed all of the relevant documents to be accessed from any occurrence of the name. This also included Bitmap drawings were textual names were added to an area within the image. To extend the set of relevant documents, all the author had to do was create another generic link with the same name in the destination document. This had the advantage that it was easy to implement and reduced the number of links required to connect all the relevant information. However, there were several drawbacks with this approach: -

• It was difficult to maintain in that using Generic links formed a primitive cluster made up of a collection of links. Therefore, to edit the cluster involved finding and change each of the links.
• The Generic link method was limited by having only one method of entry into the cluster of links. For example, if all the entries into a cluster were keyed to the name Y145. Yet at the same time Y145 was also referred to by another name. It was impossible to build a cluster that was able to encompass both representations.

Hence, a link cluster was developed that solved these problems. This was achieved by making the cluster an object with in the information system. By making the cluster objects allowed:-

• Access control lists can be attached to the clusters, allowing the author to control access by specifying an arbitrary list of permissions for users and/or groups.
• Multiple source anchors (names), some of which can be generic and hence content free, that is referenced by other clusters.
• Maintenance of the cluster to become easier.

This model of link clustering has been implemented within the administrative overlay environment. The author was able to build complex relationships between the various MHAs using high level terms represented by these link clusters. The users were able to navigate the information using the clusters, providing an additional method for locating and browsing the information space.

Back to Contents List.

2.7.7 Revision Control.

The method of documentation control and revision control will vary from company to company. Therefore, the design methodology does not aim to replace the company’s documentation management system, but works with it. However, the links themselves are sources of information as much as a line of text or an object in a drawing. Hence it is necessary for the hypermedia system, in the revision control process, to log the changes made to the application linkbases, when the document are updated. To ensure this does not become an administrative nightmare, only links that are changed were recorded.

As stated earlier the maintainability of the system is improved and revision control of the link made easier by using the hierarchical structures of the information space and authoring using MHAs. That is, as a MHA is amended only the linkbases associated with the MHA needs to be ‘frozen’ and stored as part of the change process. In the same way a copy of the document prior to the amendment is kept as part of the document change procedure. In addition, only the linkbase that have been effected by the change and not all of the linkbases associated with the MHA needs to be ‘frozen' hence, reducing even further the amount on information stored after an amendment to a MHA.

2.7.8 Hardware Selection.

No matter how well the information is presented or intuitive to use, it can quickly be undermined by the frustration felt by the user not being able to access the information quickly. This slowness of access is largely due to the poor choice of the computers to be used on the factory floor (i.e. the computer’s physical interface and hardware specification) and the computer network to be used.

The areas to be considered by the design term are:

• The environment in which the equipment is to be used. The factory floor is a hostile environment for any piece of office equipment, in particular computer-interfaces, i.e. dirt, dust, greasy fingers etc.
• The delivery of the information to the point of use. In practice the delivery computer needs to be taken to the point of use. It would be of little benefit if the shop-floor users had to walk over to a workbench to access the information. For the portable computers battery power is still a problem.
• Where the information is to be stored and accessed. That is, storing the information locally or on servers and use local area network or wireless networks. To really exploit the use of electronic information, the original documentation is to remain on its original server. Hence there is no need to keep ‘controlled copies’ of CD-ROMs, thereby removing (by not copying) the large administrative burden (and hence cost) created by controlling a paper based system.
• The ergonomics of operating equipment and using the hypermedia system.

The answers to these question will effect the hardware selection.

2.7.9 Design Reviews.

The requirements and procedure for industrial design reviews are well established. A general excepted model is the one described in BS EN ISO 9001 [BSI 94]. In addition to the general guide for this standard, there is a guide to implementation of BS EN ISO 9001 specifically for software [BSI 91]. From Figure 2-3, it is possible to go into an infinite loop, if problems are encountered at the design stage. In practice this rarely, if ever, happens due to good project management and the constraints of time and cost.

2.7.10 Prototype.

The prototype must be representative of the industry environment as a whole. This will include the factory-floor and office personnel. A practical approach is to use the information associated with a process line, or part of a process line depending on the size on the associated information space. This approach is effective, as it will include a number of personnel (operators, clerks, maintenance technicians, engineers and managers) from a number of job functions (production, maintenance, stores, quality, etc.) During the testing of the prototype, expert evaluation of the user interface takes place [Wills 97]. Thereby reducing the number of major usability problems encountered in the user trials. In addition, the training program and test scenarios are to be device at this stage in preparation for the user trials.

2.7.11 User trials.

This stage is a very interactive stage between the users, evaluators and the design team. The evaluation method used will very depending on the application domain (size and requirements). However, in an industrial environment the evaluation can be, to some extend or other, be divided into 3 mains areas:-

• Training the users. This is an important element of the trails. It is good practice to profile the users prior to training, as this is useful for assessing the effectiveness of the training program and system at a later date.
• Conduction system evaluations, this is the formal part of the user trails. Here the users perform pre-set scenarios with comments on the usability and suggestion for improvement being feedback to the design team.
• Allow the user freedom to explore the system. Here the user are let loose on the system using as much or as little as they require for their daily work. This is an essential part of the trails and can be the most constructive, especially when combined with data logging and logbooks or questionnaires (to record errors or and user option of the system.

  Back to Contents List.

  ---

3. Conclusion.

This report has reviewed the current design methodology reported in the literature. The general design solutions described above (section 2.1-2.6), assume that the authors of the hypermedia application have control over the form, size and/or the content of the information space. In an industrial environment, there already exists an established information space, often consisting of several different information systems. These will already have their structure, content and access methods. Therefore, existing information systems need to be encompassed into the design of any industrial hypermedia application. In addition, the existing models are general models, often citing examples far removed from an industrial environment. Hence, in section 2.7, the authors have set out a design methodology that takes into account the requirements and constraints imposed by the industrial environment. The design methodology in this research removes much of the generality and focus in on the requirements and considerations for designing large-scale industrial hypermedia applications. The strategy has been used to design and implement an industrial hypermedia system, which is currently under user evaluation.

Back to Contents List.

Acknowledgements.

The authors acknowledge the EPSRC (Engineering and Physical Science Research Council) for funding the work under grant number GR/L/10482.

References.

Back to Contents List.

Copyright © 1998, University of Southampton. All rights reserved.