This article appeared in the November 1989 issue of the Videodisc Monitor. Appears here by permission.
Return to Multimedia and Digital Commentary Online
There are two serious problems that face the interactive videodisc (IVD) industry today. First, cost-effective materials development techniques are needed to reduce the high up-front costs of IVD development. Second, the needs of the end-user must be met in a service-oriented manner. Meeting both of these objectives would go a long way toward improving the climate for greater expansion in the interactive videodisc industry.
This article will in general address ways to reduce IVD materials development costs. Specifically, it will show the role that graphical user interfaces (GUI's) and windowing environments can have to this end.
As I observe the state of the interactive videodisc industry, it is becoming increasingly obvious to me that there are a large number of one-of-a-kind videodisc IVD applications. A lot of folks will do one large-scale project but not many people want to do two. Many individuals (end-user types) who have put their heart and soul as well as their own blood, sweat, and tears into one project aren't sure they ever want to work on another. And many organizations who have paid for one application often aren't ready to pay for a second.
What is missing is the ability to move from the level of having each videodisc application be a unique work of art, to being something that is reproducible on a large scale, with each, all the while, remaining interesting and useful. What we need is not to produce individual Renoir or Monet paintings, but rather etchings that can be reproduced from master metal plates in large quantities, enabling many to benefit from the skill of the artist and to enjoy his work.
This analogy perhaps breaks down somewhat when applied to interactive videodisc applications, for there are few subject matter areas that will lend themselves to only one form of interactive treatment. It is critical to recognize, however, that if videodisc is to be successful, then easier ways must be found to create large numbers of creative but yet cost-effective applications.
With respect to the whole problem of combining productivity with creativity in the development process, one aspect escapes most people: there is not just an "authoring" issue that needs to be addressed. There is more to solving the problem than just giving someone an effective tool for planning an application and entering data into the computer. Generally ignored is the fact that there are really two parts to interactive videodisc materials: structure and content. Structure is the interaction with the material that the author creates for the learner, and content is the material to be conveyed.
The distinction between the two is lost on most authoring system developers, such that content and structure are usually "mish- mashed" together during the authoring process. Even if content is available from other sources in a machine-readable form, the information often must be entered into authoring systems through the menu structure that is designed to "facilitate" course development. When it is possible for structure to be carried from one lesson to another, it is often in the form of a template of the desired interaction. It remains to the developer to enter the new material or content into the template, often a very labor-intensive process.
The best analogy to explain this mode of operation is to consider how an individual might compute the average of a set of numbers using a programming language. To follow the logic of the typical template operation mentioned above, the user creates a "program" to do the necessary calculations. The program is then compiled and executed and remains useful as long as the need exists to recompute the average of the same set of numbers, an action obviously not often repeated! To calculate the average of new numbers, the program must be edited to place the new values into the proper place of the logical flow of the program. The new results are obtained after the program has been re-compiled and re-executed.
By separating content and structure in this example, great economies of effort are quickly realized. The program is made "data-driven" by using variable names in the program for the computation of the arithmetic average, with input statements being used to read in the numbers (or content of the variables) on which the math operations will be performed. Thus the need is removed for the end-user to be a programmer. Each time a new average is required, the program is executed, the new numbers are entered (obviously without having to worry about programming syntax), and the results are displayed or printed as before.
This principle is often applied in the from of menu-driven authoring systems, but unfortunately the structure that is imposed on the developer is frequently found to be too restrictive. For this principle of data-driven software to be effectively applied to the development of creative IVD materials, then a solution must be found that avoids the loss of flexibility of most menu-driven systems and the difficulty of use of programming-based systems.
Needed are powerful tools for producing the interactions and managing the data needed by them. The whole process implies a disparate set of functions (instructional design decisions, graphics editing, text editing, the creation of interactions, videodisc frame number collection and storage, as well as various other data management functions) that are each different enough from all of the others so as to present a daunting learning curve for anyone called upon to work in the development process. Systems that attempt to solve each of these problems generally become so complex as to discourage even the most motivated user.
Even if complexity of use is not too much of a problem, the attempt to create software that meets the various needs encountered in the authoring process will generally "water-down" the capabilities of one or more of the functions. The same difficulty is encountered with many "integrated" office software packages that attempt to provide the capabilities of a word- processor, database management system, spreadsheet, graphics package, and communications program. Aspects of the integrated package pale as critics compare them with dedicated software.
The problem addressed by the integrated software and by authoring systems is the same in both cases: providing a better interface for software without sacrificing functionality. Users everywhere are clamoring for powerful software that is easier to use. It appears not only that demands for improvements are being made, but also that the software people are listening. For as Jim Manzi from Lotus said at COMDEX a year or so ago, "If automobiles were like computers today, one would have to be a mechanic to drive."
The development of improved user interfaces has indeed become a major goal of the largest makers of "conventional" software, with the largest companies expending significant effort to this end. For example, both Lotus and Ashton-Tate have struggled with the newest versions of their main software products, taking what seems to some to be inordinately long periods of time to finish them. It does not seem out of line to guess that the greater portion of this effort has been spent working on making interfaces more user-friendly.
For example, it has been said that the quantity of dBase IV code increased from 50,000 to 450,000 lines! This seems incredible, given that the new software will serve essentially the same basic function as the older version. Because the function remains essentially the same, there should be no doubt that a major portion of this programming effort went toward improving the user interface.
Of course the objective is understandable, but given the development delays being experienced, one has to wonder if the solutions being pursued are the best available. Both companies' efforts no doubt involve considerable work being expended on developing a better user interface for each of their software products. Almost by definition this would imply a certain redundancy on the part of both programming teams: some of the same problems are being solved twice.
Proponents of Graphical User Interfaces (GUI's) maintain that such duplication of effort is unnecessary. They claim that putting software under a GUI environment takes care of a significant portion of the work that would otherwise be spent on creating the interface. According to them, it is necessary only to call the subroutines that create the desired interface. (Not usually a trivial process as will be discussed later.) In addition, by working in this way, the developer constrains the software to a certain style, creating a uniformity of interface that helps the end-user move from one package to another, addressing the steep learning curve mentioned earlier. Thus, Graphical User Interfaces work within and with a particular windowing environment to provide a standard "look and feel" across applications.
This approach brings the user a level of functionality that is very difficult to achieve with more conventional software. When working in such an environment, the user is less concerned with remembering arcane syntax or command sequences and thus can concentrate on more significant creative issues.
This advantage, along with several other potential benefits, will be discussed below. Let it suffice for the moment to say that GUI's appear to be an important key to the software industry's ability to address the need to improve user interfaces, let's look briefly at their origin and development. We will then examine the advantages of these environments and relate these to IVD materials production. Next, we will discuss briefly the negative side of developing the applications to run under windowing environments and show how the problems are being solved. Finally, we will end with a discussion of where we think windowing is going in the future.
An interesting date for Graphical User Interfaces and windowing environments is Super Bowl Sunday, January 1984. In the third quarter of that game, as the Los Angeles Raiders pulled ahead of the Washington Redskins, Apple Computer ran its controversial, $1.6 million commercial that announced their new Macintosh computer. Some stockholders wrote letters complaining of fiduciary irresponsibility, and advertising moguls harped that the commercial had nothing to do with the product. Nevertheless, it went on to win the Grand Prix at Cannes.
Aside from its artistic nature, the significance of this commercial is the message it conveyed to the world, "Apple was coming to liberate computer users everywhere from the enslaving interface that belonged to MS-DOS." The commercial showed a procession of shaved-headed, emaciated individuals in drab, baggy clothes passing in cadence through a glass-like tube into a large hall where they seated themselves before a monitor that showed the all-knowing image of Big Brother. They were then bombarded with a Newspeak discourse delivered by an authoritative narrator with the text of the message appearing at the bottom of the screen. The narrator in essence reminded them of how great it was that they had given up their individuality for the good of society.
Suddenly, an attractive, athletic woman, dressed in bright red shorts, running shoes, and a Macintosh T-shirt and carrying a sledgehammer, came running into the hall. She approached the screen, spun the hammer over her head, and hurled it into the screen, causing it to explode in a flash of light. The crowd sat with their mouths open. Mesmerized. A voice was heard, "On January 24th, Apple Computer will introduce Macintosh... 1984 won't be like 1984."
Now, some five years later, Apple Computer has of course not buried the collective MS-DOS world composed of IBM's PC family and all of its clone cousins from numerous companies. This is obvious, given the installed base of some 28 million MS-DOS machines as compared to Apple's 2 million Macintoshes.
What has happened, however, is that the graphical user interface heralded first by Xerox's Star, then Apple's Lisa, and finally by Macintosh is now entering a phase of rapid expansion on all personal computer platforms. Today, whether we consider the implications of Motif from the Open Software Foundation, the X- Window standard from MIT's Athena project, Microsoft Windows, OS/2 and Presentation Manager, the NeXT interface or any other proprietary windowing interfaces on any number of other platforms, or of course Apple's Macintosh, it is obvious that users are in fact demanding to be liberated from the rigors of the "command line interface."
What is the rationale in assuming that these new operating environments can meet this need? I first spoke of the advantages of windowing interfaces in a talk in New York in November, 1987, at the OIS '87 Conference. There I discussed research we have conducted at the Air Force Academy for the Human Systems Division of the Air Force Systems Command. Our work did not constitute an exhaustive look at authoring systems but centered on authoring system specifications that make various systems more productive than others. During our research we identified several solid rationales for using software running under windowing environments:
In our study we surveyed the capabilities of several representative authoring systems. But the "acid test" was to have experienced computer users work with the various packages and their documentation to see how long it took to learn to do something productive with the various systems. We concluded that applications running in windowing environments are easier to learn and to use.
Another advantage, transferability of skills, quickly became apparent. As we worked with the MS-Windows environment, it became apparent that once a user was somewhat familiar with the Windows interface, it was very easy to move from graphics editors, to the video editor, to text editors, without any significant learning curve each time. This was true whether the tools were part of the authoring package or were external applications running under Windows.
Just as GUI's allow users to move from one application to another, so do they allow text and graphics to be moved from application to application. This advantage, perhaps best described as "open software architecture", is extremely powerful and accounts for a good deal of the power that GUI's offer. Essentially, this means that users can integrate into their work environment the very best tools for the job at hand.
These advantages apply not only to software in general but also to IVD software in particular. Because of the large number of MS-DOS machines being used in the interactive videodisc world, the GUI phenomenon must be related to Microsoft Windows (MS- Windows) and Presentation Manager (PM). For the IVD industry as a whole, these systems have significant implications for courseware development productivity. Just as it is estimated that information workers only take advantage of 20 to 40 percent of the capabilities of their favorite software tools, so are videodisc developers certainly less productive than is ultimately possible. And just as MS-Windows/PM promise to increase productivity in the MS-DOS office world, so can they help increase the productivity of developers as they create IVD applications for MS-DOS delivery platforms.
With software that is easier to learn and to use, the developer can concentrate on more significant issues such as the design of the interactions the learner will have with the IVD system. Rather than being forgotten upon disuse, tools that are used infrequently remain within easy practical access for the developer to call upon when required.
And when current tools do not meet current needs, new ones can be easily learned and integrated into the work environment, providing the most suitable software for each part of the development process. The open architecture that is found in windowing environments fosters a collaboration between processes that is difficult or impossible to obtain with conventional software. Authoring tools, graphics software, font editors, database management systems, and text editors, each selected to best solve the problem at hand, come together for the IVD materials developer like hammers and chisels for the sculptor or brushes and palette knives for the artist. Like the tools of any master craftsman, each is carefully chosen to serve a specific purpose.
All of this is in theory quite exciting, but in practice, the ultimate, complete and adequate set of tools does not yet exist. The interested observer admittedly has to demonstrate some enlightened imagination to project what the ideal environment can be. When I first started talking about the benefits of GUI's, the advantages were not always readily apparent, due to the limitations of Microsoft's first generation of Windows that we were using at the time. There were bugs and other problems that made software development for the environment a difficult process, significantly reducing the number of applications that it could support. Thus, the requisite excitement and functionality were not sufficient to make the first version a serious contender for software developers' attention. And of course, without applications, users were obviously not interested in launching into any serious use of the system.
It is still today a very difficult process to create applications for MS-Windows (or for Apple's Macintosh for that matter), a fact that has not gone unnoticed by none other than Steve Jobs. If his NeXT machine doesn't succeed in other category, it is at least setting the standard for how applications will be produced in the future. With virtually every machine the company is shipping being a development system, Jobs is supplying formidable tools to ensure that exciting software will be forthcoming. As an example, NeXT's Interface Builder reduces the repertoire of subroutines that must be mastered by the developer's from several hundred for creating Macintosh or Windows interfaces to only 34, helping NeXT meet their goal of reversing the ratio of time spent on interfaces vs that spent on programming the functionality of the typical software package that will run in a windowing environment.
While NeXT seems to be setting the pace, things do seem to be getting easier for the MS-DOS world, thanks to several new developments in the MS-Windows software arena. Paul Allen (Bill Gates' partner in the creation of Microsoft, and still the company's second largest stockholder) and his new company, Asymetrix, are developing a Hypercard-like applications generator for the Windows environment. Gupta Technologies is in their second release of SQL Windows, a tool for creating Windows-based interfaces for accessing SQL databases. The Whitewater Group has produced the object-oriented programming environment, ACTOR. And finally, a California company has released dBWindows, a product for creating Windows applications that access dBase III files.
Furthermore, it appears that things are just getting going. Not only is there a growing sophistication in the tools that are available, but also programmers are moving swiftly up the learning curve for producing applications. Where in recent months there was an apparent shortage of qualified Windows programmers, today the number seems perhaps to be reaching a critical mass, somewhat reducing the hiring difficulties of software houses.
But not only is it just getting easier to produce applications. The capabilities of the environment itself are on a definite rise. Already, versions 2.03 and 2.11 brought significant increases in performance and features. For example, 2.11 enabled access to an additional 64K in high memory, somewhat relaxing memory constraints and increasing the performance of large software packages.
As reported in the trade press, Windows 3.0 is due out soon with much-improved memory management that is surprising Beta users. Reports state that most or all of the system has been recoded in assembly language rather than in C, significantly reducing size while increasing the system's speed. There are also indications this version will contain animation intrinsics, a capability that will greatly facilitate the development and execution of CD- ROM/XA applications.
Finally, there is Hewlett-Packard's additional layer of capability that functions within Windows: NewWave. According to an Infoworld column last Spring by Jerry Pournelle, it is NewWave that has caused Apple to sue Microsoft and HP. Impressed by what he saw in early demonstrations, Pournelle's conclusion was that with the improvement NewWave brings to Windows, Apple has every right to be concerned. American Airlines has been impressed enough with NewWave to order its implementation throughout their vast network of mainframes and microcomputers.
Success, however, is not absolutely guaranteed by technical quality. It is well-known that there are a host of other market- related phenomena that determine a product's acceptance by consumers. Perhaps the prime factor that will determine the success of any hardware or operating system is certainly the question of software availability. A lot of people acknowledge that Sony's Beta system was technically superior to VHS, the winner of that fierce marketing battle. While the unit's selling price was certainly an issue, VHS technology allowed for more material to be recorded for less money, significantly reducing the consumers' software costs as they created their own libraries of materials. In the same way the battle for market share for VCR's was tremendously influenced by software availability, so will success of windowing environments be determined by both the quantity and quality of applications available.
For example, to date, spreadsheets and word processors have sold computers. Visicalc sold Apple II's and Lotus 1,2,3 sold IBM PC's. Today, on software best-seller lists everywhere, Microsoft Excel (running under Windows), is second only to non-Windows product, Lotus 1,2,3. Samna's Ami word processing software and their soon-to-be-released professional version combined with Microsoft's own Word for Windows will complete the requirement for word processing software. In addition, there is a wide-range of graphics software already available from several vendors.
Just as Windows 2.03 and 2.11 have turned this situation around, Version 3.0 promises to continue to dispel the misgivings of developers. There is currently a flow of interesting applications into the marketplace that shows no sign of letting up. On the contrary, activity is increasing.
Finally, there is the issue of Macintosh vs Windows. Considering there are about 15 times more MS-DOS machines installed than Mac's, there is a level of inertia that has to be taken into account. While business sees the advantages of the Macintosh, they are not ready to throw out their installed base, and thus are showing an increasing interest in MS-Windows.
In summary, there are a lot of reasons not to underestimate MS- Windows' potential for success. Whatever the reason and probably to their own surprise nevertheless, Microsoft has been shipping 50,000 to 70,000 copies of Windows a month for some time now. Windows 3.0 will no doubt be included with many machines being shipped at year's end and there are reports that it will even be pre-installed on a lot of hard disks when they leave the computer manufacturers' factories.
Subsequent to our research group's conclusion of the importance of Microsoft Windows, IBM and Microsoft announced OS/2 and Presentation Manager (PM), with shipment for the latter finally taking place on schedule at the end of October, 1988. While there are few applications available to date that take advantage of this system and nay-sayers on the importance of this system abound, one must remember that Microsoft has sold at least 5,000 development kits for PM-OS/2 at a cost of $3,000 each. Fifteen million dollars represents an enormous investment for the microcomputer software industry and when combined with the costs of the programmers who will use the kits, it is apparent that developers are making a significant commitment to this software environment.
It is true that memory costs and other hardware requirements have slowed PM's acceptance, but I am convinced that the long term picture is bright. Just as software was developed for MS-DOS that was more powerful than was possible under CP/M, so will the rich OS/2 and Presentation Manager environment foster applications that cannot be run under MS-DOS. At that time, users will surely decide that the functionality of the software justifies the additional investment required in hardware.
One of the things we know from experience in general and from OS/2 in particular is that software will expand to fill machine cycles and memory available. The positive side of this "law" is that this computational potential will be put to work creating software that is significantly more functional. The potential also exists at the same time for the software to be made more user-friendly.
Two examples of the additional functionality that will come from these new operating environments are: (1) Interprocess communications such as Dynamic Data Exchange (DDE) under MS- Windows and (2) true multiprocessing capabilities such as client/server architectures on a network. Both of these promising developments are made possible by object-oriented programming principles that are all but inherent in windowing environments. The applications these capabilities make possible are truly exciting for work-group computing in the IVD creative process.
I feel very strongly that the adoption of Graphical User Interfaces into the workplace is inevitable. The less than desirable productivity rates of information workers using computers, combined with what is in reality a low degree of penetration of machines into the workplace, are both evidence to the information industry that a change is warranted.
As far as concerns the IVD materials development business, I believe that Microsoft Windows and Presentation Manager, the primary windowing environments for MS-DOS machines, are critical elements that can help bring significantly increased productivity to the interactive videodisc industry. The view that a change is coming was supported by Joe Larsen, President of ITC, in his presentation to the Society for Applied Learning Technology's (SALT) Conference in Orlando in February of 1988. Stated simply, Joe's opinion was that in the near future, software for interactive training applications that does not run under Windows will not be able to compete in the interactive training marketplace.
But these environments are only a part of the picture, an extremely important part, but still only a part. It is up to software providers within the interactive videodisc industry to work to build the productivity tools to achieve the maximum benefit that these windowing environments can offer. This benefit translates into significant increases in productivity for developers as they produce materials. What is good for IVD materials developers has to be good for the industry!