The Executive Information System (EIS) project is a pilot effort by the Information Systems Division (ISD) to use "system visualization" to develop a critical internal information system. The project is transferring technology which was pioneered by multi-year internal research on improving the system and software engineering development cycle.
The EIS Project was started in 1990 as a side effort to the Hughes Space and Communication's Group's (SCG's) $50M overall information systems modernization effort under Gary Osborn, chief information officer. The initial focus was to permit the Group President to access operations status and Group metrics in an easier, quicker, and more consistent manner using modern desktop technology.
Use of the EIS by Group Vice Presidents, Division Managers, Program Managers, and key staff insures everyone reads from the same "sheet of music." Reports are available on all systems and operations divisions, subsidiaries, programs, and directorates.
The EIS presents highly summarized and synopsized information to executives (versus the "nuts and bolts" data typically stored by mainframe applications). Its basic metaphor is the "early warning" indicator. These indicators are organized into "traffic light" views of the various organizations. Computer screen layouts use the standard reports approved for the Group. Content for the reports comes from current data providing organizations, which now produce over 500 reports on desktop microcomputers each month (using, for example, Lotus or Excel spreadsheets and Harvard Graphics or Macintosh drawings).
The EIS team has developed a Macintosh-based prototype system using commercial off-the-shelf products and some custom software. The EIS prototype has been demonstrated to some SCG executives who are likely users. The go-ahead was given to deploy a pilot system for 5 to 10 executives.
The main issues affecting 'buy-in' of data provider organizations and executives are scope, value-added above current operations, data security, data timeliness, burden (cost) on providers' staffs, and performance.
System visualization is a suite of process technologies and techniques used to enhance the system engineering and design efforts prior to the software manufacturing phase. Rapid prototyping is but one technique. A particularly effective technique for communicating with very busy and dispersed customer representatives has been recording computer screen and live video material onto the 'laser printer of multimedia', namely, the VCR. Electronic co-location permits 'live' sharing and interaction.
Visualization for software systems applies and expands approaches which have long been successfully used in other engineering disciplines. Examples of such use are architectural models, stereolithography, automotive 'clay' models, and spacecraft mockups and work habitats. System visualization avoids one of the traditional major problems with development of software systems, namely, long intervals of "magic and mystery" due to the invisible nature of software: you do not see what you will get until it's built. Visualization makes tangible and concrete to customers and end-users the look and feel and functionality of their future system and how they will interact with it. That is, this approach models the essential operational interfaces and procedures, namely, the human, technical, and organizational aspects of the customer's work environment. Visualization mitigates both cost and technological risk.
The ISD has invested heavily in maintaining itself at the forefront of applying structured analysis and computer-aided software engineering (CASE) tools. These methods and tools offer a disciplined approach to generating the contractor's 'blueprint' and reducing manufacturing and operations and maintenance costs. This blueprint, however, does not necessarily communicate well to customers and users. Structured methods and products do not, by themselves, adequately facilitate critical conceptual and proof-of-concept efforts. System visualization facilitates system engineering activities in answering the basic questions:
System visualization incorporates insights into how people discover and close on the particulars of complex problems and designs. "I don't know I want but I'll know it when I see it." System visualization does not focus on "speeds and feeds", but rather the frame of reference, the task focus, of customers and users. Customers and users participate from the start.
System visualization was not very practical for software systems until recent advances in desktop technology and multimedia. The cost of an agile desktop multimedia toolkit is now within reach of typical business campaigns. The Apple Macintosh currently is the most cost-effective platform on which to integrate the inherently graphical visualization toolkit. In a matter of hours via real-time interactive (and iterative) focus sessions, customers and users can actually get 'hands on' the shape of their future system. The 'quick and dirty' models quickly stimulate and provoke reactions, elicit unconscious predispositions and assumptions, and sanity-check operational constraints. The process rapidly bootstraps to more complete disclosure of expectations and packaging for the new system. Engineers and customers communicate using a shared model.
Central to the visualization approach is careful management of different decision points. Since the focus is not on "speeds and feeds" (although critical-path technology demo's are slotted in where appropriate), mockups and working models (proof of concept), may not satisfy certain requirements. Furthermore, the visualization toolkit may not be appropriate for the next phases of prototyping and production. These distinctions need to be shared by everyone involved. The path is from being strawman and illustrative to being realistic to being prototypical to being real. Prototypes, for example, get closer to the build-to design, which may use different technology and/or platforms, depending on satisfaction of evolving critical requirements and budgets. But all along the way the customer is exposed to 'live' demonstrations of what will be delivered.
By maintaining an early focus on faithful rendering of the customer's business, in a dramatic microcosm or habitat, the complete suite of operational characteristics -- the human, technical, and organizational aspects of the customer's work environment -- will be achieved.
The EIS will be used by executives on a daily basis in order to:
The EIS supports the general goal of weaning the Group away from paper output.
These goals pose unusual challenges in that:
Executives expressed a high degree of interest in the following matters:
Data-provider organizations expressed a high degree of concern on the following matters:
As a result of the above concerns, the baseline system was designed to minimize development and operations costs as follows:
Whether or not the current EIS prototype is widely deployed with the Group, the project is interesting from several points of view:
For some time there has been a working prototype/model. This Macintosh-based system was demonstrated, for example, at the ISD Institute's 'Object-Oriented Tools' seminar in July 1991.
Work on the EIS has been an opportunity to transfer technology which was pioneered on past process technology projects. Development confirmed repeatedly that mockups and visualizations, done with an agile toolset in real time focus sessions, are the best way to stimulate customers and users into clarifying system scope and concepts, requirements, and user-interface design. More importantly, the effort has consolidated procedures to do such discovery.
System visualization incorporates not only insights from cognitive psychology, learning theory, psychodrama, group dynamics, and system modeling and simulation, but also recognizes current business dynamics and problematical budgets.
The very improvements in contractor internal development methodology and discipline, e.g., our rising level on the SEI assessment scale, produce more and more complex and lengthy technical specifications (in paper or softcopy form) which customers and users increasingly have limited ability or time to study and grasp (from their perspective). What has been missing in system engineering efforts is strong facilitation of a shared (system) vision which permits accurate and rapid closure on key concepts, requirements and high-level design.
System visualization consolidates a visual statement of system requirements. It is particularly useful for systems which have no precedent, have not been created before, and whose success relies heavily on continuous clarification of fuzzy user expectations throughout the engineering cycle. It is also particularly useful for systems which span a large number of varied organizations which may not share a common vision of the system. Advances in the last few years in software tools now make it possible to apply visualization techniques using even small microcomputers like the Macintosh.