oturn home > Theory of presentation: overview > Appendix I-Cinegram

A theory of presentation and its implications for the design of online technical documentation
©1997 Detlev Fischer, Coventry University, VIDe (Visual and Information Design) centre

Appendix I-Cinegram

This appendix describes the cinegram prototype in outline. It is mainly provided for those readers who lack the facilities to explore the cinegram interactively. Implementation details are neglected.

The starting point for the cinegram design was the document type of Engineering Technical Graphic (ETG) with its horizontal diagrammatic integration and richness of optimised detail. (cf. App. III-Engineering Technical Graphics). An on-line document would allow the incorporation of dynamic and temporal system aspects and provide pointers to related information stored in a network of nodes.

The initial decision to base the cinegram on the ETG of the oil system of one of Rolls Royce's most recent turbine engines was taken after a week of interviews with a number of people in different departments at Rolls Royce. The decision was primarily guided by the recognition of the great popularity and multi-purpose use of ETGs in many different domains within (and outside) Rolls Royce. The cinegram was therefore not defined for a single -purpose scenario of use, but rather as a resource with different instantiations for referencing and learning (both exploratory and classroom-based).

The cinegram prototype was built and refined over the period of 2 years. Research in the field, design, implementation, and formative evaluation went hand in hand. The structure and intended scope of the cinegram prototype changed many times during the research. Instead of an all-inclusive history of these changes, I will only summarise the development.

The development began with the concept of a relatively self-contained application limited to the oil system, and developed towards a modular architecture where other cinegrams covering other ATA chapters might be added. At the beginning, there were few concerns about customisation, updating, and interfacing to other systems like service history databases. Such ideas emerged while the author spent time in the field learning more about the work context, its presentations, and the role of documentation within it, while continuing work on the prototype. This learning process found expression in several structural re-designs of the prototype when it appeared that a chosen architecture had been too limited.

Spatial architecture

The cinegram prototype was implemented in Apple's HyperCard™. Contrary to industrial-strength hypertext applications, HyperCard is more often used in and for educational settings [1]. It is regarded as ‘rather simple’ (Rheinsperger 1992, p14), ‘not truly object-oriented’ (Kay 1993), and accused of imposing a ‘tyranny of buttons’ (Hall 1994). HyperCard was one of the first commercially available Hypermedia applications which consequently spawned a number of structurally similar applications such as SuperCard, Toolbook and Coursebuilder.

The author assumes broad familiarity with the notion of card-based hypermedia authoring software containing scripting languages. For the reader interested in implementation details, there exists a plethora of books on multimedia tools and techniques, including a range of books on HyperCard.

Overview diagram

The first view (or root node) that appears when launching the cinegram is here called the document midpoint or simply midpoint, to indicate that it is the central anchor to which users often return during navigation. In terms of the hierarchical structure of the system, the midpoint is the equivalent of the list of contents of a chapter within the maintenance manual, e.g. 79-Oil system (cf. figure I.1). This root node could be just one of many views in a superior structure, which may, for example, contain all ATA chapters referring to the engine.

ATA chapter structure

Figure I.1. The cinegram prototype is based on one of the 10 ATA chapters devoted to the engine, 79-Oil system. Three stacks provide different access structures depending on the relevant domain: Description, as a general overview for training purposes; Troubleshooting, where the top level presents common fault indications; and Maintenance, focusing on location drawings and photographs, where installation and removal sequences can be presented for arbitrary access or linear presentation. In the current prototype , only Description is fully implemented. The figure shows the main overview diagram and the links afforded to both subsystem views and straight to component views. Importantly the structure includes components such as the Centrifugal breather (bottom row left) which in terms of ATA classification, belong to another chapter.

On launching the cinegram the document midpoint calls up an animated overview diagram, a navigation window, and a text window (see figure I.2). The chosen screen real estate makes it impossible to show components in as much detail as is employed in simplified A3 ETGs. A reduction in both token size and internal complexity was necessary. However, while icons for pumps, filters and coolers are rather small and schematic, more size was allocated, and more detail retained, in bearing chamber and gearbox tokens. By virtue of gestalt laws, the centrally aligned bearing chambers can be seen to hold the three rotating assemblies (shafts and drums) rendered in different shades of grey. The level of detail also affords a view of the number and position of oil jets within bearing chambers and gearboxes. 

Animated overview
diagram of the cinegram

Figure I.2. The illustration shows the animated overview diagram of the cinegram (description mode). The icons show all system components and simultaneously act as pointers to the respective default component view. The animation shows the circular oil flow through all components; the oil feed pipes are shown in red, the scavenge (or return) lines are shown in green. The pipes venting oil-air mist from the bearing chambers are shown in a light yellow-green.

The layout of the overview diagram shows a clear departure from the topographical design of ETG overview diagrams (see figure I.3). In the cinegram, the clear layout of the flow topology took precedence over topographical location. The diagram is roughly divided into three horizontal areas: feed oil in the upper third, lubrication in the middle, and scavenging in the lower third. This grouping is reinforced through colour-coding in line with Rolls Royce's conventions. Animation provides an immediate cue as to the direction of the flow. Overall, the oil circulates in clockwise direction. The redesign has reduced the number of pipe crossings from 51 to 16 (less than one third of the original number). This has clearly reduced visual complexity of the layout.

topological comparison of the simplified oil system ETG (a) and the cinegram overview diagram

                         (a)                                                         (b)

Figure I.3. The illustration shows a topological comparison of the simplified oil system ETG (a) and the cinegram overview diagram (b). In both views, components are numbered in the order of oil flow. The ETG retains some topographical fidelity: a number of components (3,4,5,9,10, and 11) are shown close together. The rationale behind this is the fact that all these components are part of the same assembly (or ‘lump’). The disadvantage is that this leads to much more convoluted pipe connections, making it difficult to follow the oil flow. The cinegram overview diagram sacrifices topographic fidelity for increased functional clarity. The number of pipe crossings is reduced from 51 to 16.

Navigation window

While the display of composite media is invoked when the user moves to a particular node, there is also a constantly available navigation window at the top-right side of the screen. It provides generic access to nodes via a pop-up menus on the three levels of the hierarchical ATA classification (for an explanation of the ATA classification, see App. II-ATA-specified documents). Accessing a new node will update the navigation window so that it reflects the current position within the hierarchy of chapter-section-subject. The navigation window also indicates additional locally relevant material, e.g., other views of a component, appropriate animation schedules, etc.

Component views

In the cinegram version, the components which appear on one uninterrupted surface in the ETG are distributed to different nodes. Any mouse selection over one of the overview icons switches to a component view which calls up a different set of composite elements. (see figure I.4).

scavenge filterScavenge pumps

             (a)                                                 (b)

Figure I.4. The illustration shows two topologically adjacent ‘lumps’ each of which consists of several components. The currently selected components are scavenge filter (a) and Scavenge pumps (b). Physically, the lumps are connected by a ca. 6 feet long combined scavenge return pipe (the connection is indicated through the dotted line). In the cinegram the links to adjacent components can be followed by clicking at the yellow labels at oil in- and outlets.

In ETGs, there is a structural separation of the system into lumps of materially adjacent components which are shown in optimised cross section views. This separation informed the cinegram segmentation of the Trent 700 oil system into discrete nodes. The scavenge filter view, for example, carries the main chip detector in its inlet, a ÆP switch measuring the pressure drop across the filter, and two temperature sensors in it outlet. In the description mode, the default view type for all components is a cross-section animation. The same  is used as the basis for slightly different animations for the different components (and component states). Figure I.4 shows two adjacent lump templates which each show a number of different components.

Each component view is based on scanned and edited ETG cross-sections. Using these views that were already being designed for the printed version shortened the time spent on drawing and ensured a standard of technical drawing skills which was beyond the expertise of the current author. Also, the similarity of component views in cinegram and printed ETG would afford easy recognition. It turned out that engineers had indeed no problems working out the cinegram views.

Several devices are used to differentiate the component animations based on the same lump template. Apart from updating the title at the bottom of the display, the text in the text window, and the pop-up menus in the navigation window, the yellow pointers placed at oil in-and outlets change to indicate standard affordances (cf. figure I.5 a-d).

Different generic pointer types used on component animations

             (a)                    (b)                     (c)                      (d)

Figure I.5. The illustration shows the different generic pointer types used on component animations (cf. also figure I.4).

The currently selected component has its pointer framed in black (b). In correspondence to this change, pointers to components that are included in the same lump animation are rendered in normal style without outline (c). Pointers to components outside the current lump template show a small arrow and italicised text (d). A zoom-in effect centring on this label on entering the component node reinforces the selection (a).

Temporal architecture

There are three different types of dynamic presentation in cinegrams: steady-state animation, state changes, and schedules composed out of any number of cinegram nodes.

Steady-state animation

Steady-state animation is realised through short looping animation segments called up at a single cinegram node. The animation is based on a refined version of the . An analogue variant exploiting the movement of interference patterns generated through a large rotating wheel behind a partially transparent display has long been used for control displays in power plants (cf. Kieras 1992).

Steady-state animation affords the differentiation of a number of system dimensions which would require symbolic coding in static displays. It differentiates between static and moving parts and thereby facilitates part boundary detection. It shows flow type, flow direction, and flow speed as continuous travelling wave pattern. Speed, contrast and gradual changes of colour can denote changes in operating conditions.

State changes

The second type, state changes, links two steady-state animations via one or more  segments. Each transitional segment calls up the following segment on reaching its end, until the target node is reached. State changes can reflect different kinds of activity, such as operational changes, exceptional changes (such as filter blockages), changes caused through maintenance operations, or changes redirecting the focus of attention to particular parts or aspects of equipment or processes. In each case, animation can bridge the gap between states through transitions.

The following example of state change illustrates the behaviour of a system component during engine operation. The pressure filter has a containment valve that prevents oil from leaking out when the filter element is removed. Figure I.6 (a) shows how different states of moveable parts are usually shown in engineering schematics: a vertical line divides the filter and shows the open and closed position of the containment valve on either side. Even for an experienced engineer, it takes some time to work out the arrangement of the nested valves. The animation—which may be imagined as a smooth transition between Figure I.6 (b) and (c)—immediately shows how the valve responds to a filter removal. When the engine is shut down, the small spring-loaded inner valve closes off the oil supply from downstream of the filter (b); on unscrewing the filter housing, the entire spring-loaded containment valve is brought down to seal off the oil from upstream of the filter (c).

small segment of one schematic cross-section of the pumps and filter housing assembly

             (a)                               (b)                                      (c)

Figure I.6. View (a) shows a small segment of one schematic cross-section of the pumps and filter housing assembly. Two states of the automatic oil containment valve are shown to either side of a vertical dividing line. View (b) and (c) show two states from a filter removal animation schedule.

Schedules

Schedules in the cinegram combine any number of animated cinegram nodes as dependent segments. On reaching its end, an animated segment triggers navigation to the next node of the schedule which will transparently call up the next segment. All segments are listed in a  which allows playing schedules as a whole or in single steps. Any segment can also be accessed arbitrarily (cf. figure I.7).

The same view can appear in different schedules under a different segment name. Schedules were limited to the individual component level in version 2 of the cinegram; in version 3, they can span across the entire ATA structure, transparently linking views which technically belong to different systems.

cinegram component view with the schedule control
panel opened

Figure I.7. A cinegram component view with the schedule control panel opened in place of the navigation window. Controls afford playing or stepping through schedules. The schedule has reached the fourth segment, which is reflected in the highlighting of the fourth list item.


Footnotes to Appendix I-Cinegram

[1] One example of an industrial application written in HyperCard is the FAKS system described by Priha (1991).

Last update: 08 November 2007 | Impressum—Imprint