|Building Blocks of Web Portals|
|How to Master Portal Design|
By Gerd Waloszek, SAP AG, SAP User Experience – August 1, 2000
Disclaimer: Please note that this edition was written in 2000. Therefore, statements in the articles, particularly those regarding SAP's products, product strategy, branding strategy, and organizational structure, may no longer be valid.
In this article we put portals into an historical perspective with hypertext, the Web, and the desktop metaphor, each of them an important step in the direction of making computers universal information management machines. We finally ask whether portals are really a step beyond the desktop.
When Vannevar Bush invented the hypertext principle in the 1940s, computers were just emerging. So, Bush envisioned his MEMEX system with microfilm as the principal medium for communication. Twenty years later, when Ted Nelson coined the term "hypertext" and started his influential XANADU, a world-wide network of linked documents, computers were a reality and formed the technological basis for hypertext systems. Nelson was followed by many others who created more or less sophisticated hypertext systems - ranging from Ben Shneiderman's simple DOS-based HyperTies to large-scale systems like the Xerox's NoteCards or CMU's KMS (Knowledge Management System). But no matter how sophisticated these systems were, they were typically restricted to one computer or only small computer networks. Large-scale, or even global computer networks like the Internet, were just emerging - these systems did not take advantage of them.
When Tim Berners-Lee, a scientist at CERN (Centre Européen pour la Recherche Nucléaire) in Geneva, Switzerland invented the World Wide Web (WWW) in 1989, the computer infrastructure was far more developed. Computers had become omnipresent, and the global Internet was already at their disposal. Berners-Lee conceived the WWW as an easy-to-use interface for exchanging scientific research documents throughout the whole world. To make this happen, it still took a few years, but it is reality today. The WWW may be similar in spirit to the ideas of Bush and Nelson, but it surely is a much more chaotic, or even anarchic system than either of those pioneers had in mind. The hypertext formalism developed for the WWW, however, is primitive: it has just one link type and a simple universal address system, the URL (uniform resource locator). It may be, however, that this simplicity is one of the reasons for the Web's success.
In the seventies the desktop metaphor came to life and changed the way we interact with computers. Up to then, computers were basically seen as "number crunchers" or storage systems. Then engineers took the idea of computers as "universal machines" seriously and used computers for "trivial office tasks" like word processing and business graphics. Ideas like the desktop metaphor and the WYSIWYG (what you see is what you get) principle were not only born, but also realized - even though the computers available then performed miserably, and the hardware was far too expensive. While the early XEROX Star and Apple Lisa failed economically, their ideas were carried on to commercial success with the Apple Macintosh, and even more so with Microsoft Windows. Other systems followed this approach, but none of them really challenged the desktop metaphor itself - it will surely stay with us for many years.
The computer desktop metaphor is based on the office environment. Curiously enough, this metaphor has been put forward as a general metaphor for computer workplaces, even though it is plagued with problems and limitations. Maybe many of the tasks carried out with computers still can be somehow "squeezed" into the idea of being an office task - obviously, nobody dares to change this. Anyway, now we have windows, trash cans, menu bars, task bars and the like at our disposal on our virtual computer desktops to manage the chaos which has accumulated on our hard disks: applications, text documents, graphics, mails, to name just a few things. But is this metaphor really suited to managing information and mastering functionality that users need to perform their daily tasks efficiently?
I personally believe that the WYSIWYG principle is an even more important concept than the desktop. It stands for "what you see is what you get" and it means that word processors displayed text on the screen as it would be printed - a rather revolutionary idea when it was new. But if one extends this notion of WYSIWYG to stand for a "virtual reality" on the screen, then the screen presents an image of reality, which may be hard to understand for many people: It shares characteristics with the physical world, but also introduces new behaviors that physical objects do not have - even though the objects on the screen may look like "real" objects.
It is unlikely that the inventors of the WYSIWYG principle really had this aspect of "virtuality" in mind. Actually, they tried hard to make their virtual world as similar to the real world as possible. We still find lots of software in this vain, like CD controller programs that mimic a physical CD player. But this sort of software falls short of the advantages of a "virtual" system. Virtual means that you can do things that are impossible in reality. For example, a screen may contain text and pushbuttons. If you click a pushbutton, you can change the text, display another page, etc. Can you do this with a book? Does a book have pushbuttons? Of course not - the reality that computer screens present does not have a counterpart in real life. Computers present "illusions" - illusions which we have become used to and which are part of our daily lives.
Interaction with computers has changed dramatically in their short history. User interfaces started out as switchboards and cables, followed by teletypes, early typewriter-like printers. Their line-oriented interaction style was transferred to CRTs (cathode ray tubes), but soon terminals were enhanced through direct screen addressing, which allowed characters to be drawn anywhere on the screen. Finally, an interface was developed in which graphic screens - in contrast to the character-oriented screens used then - were "misused" for plain human-computer interaction instead of for displaying charts - this brought us the revolution of the desktop metaphor. Interaction became more and more direct and felt "physical," as Ben Shneiderman's term "direct manipulation" implies. Nonetheless, the image of the reality on the screens remains "virtual" - an illusion on phosphor or LCD pixels.
One of the biggest constraints of screens is their limited size - they are just a peephole into a "virtual reality." Eventually people wanted to perform multiple tasks simultaneously on their computers and so the problem of displaying information in parallel arose. Parallel windows became the dominant paradigm - in different flavors though: overlapping, tiled, one at a time, etc. Apple Computers even tried to protect its invention of overlapping windows and sued the competition - without success, as we all know. Window handling still presents a problem to many users; there are quite a lot of control mechanisms available - the Windows task bar is one of them. Other designers like Allan Cooper favor tiled windows, but have to deal with the problem of small tiles. Show/hide mechanisms attempt to solve this problem, but they also make interaction more complex.
Portals are a new development that has its roots in web search engines like Yahoo!. Despite their short history, many terms have been invented for portals like vertical portals (or vortals), consumer portals, or business portals. Portals attempt to provide access to information and - as does the mySAP.com Workplace - functionality for certain users in a specific context. A corporate portal, for example, may offer access to all the information that is related to the respective company, its products, organization, people, relevant research, etc. Consumer portals like Yahoo! offer information which anybody may need: news, information about sports, health, business, and weather. Portals build on all the pieces just mentioned to form a huge hyperlinked information pool. This pool may be part of the Web, as consumer portals are. Consumer portals try to attract surfers by presenting themselves as a well-structured information oasis compared to the chaotic environment of the rest of the Web. Corporate portals can - as in the case of Intranets - just use web technology, but are confined to the limits of a company. Companies may, however, also give their employees access to the Web, thus potentially making their internal information pool unlimited.
Portal advocates claim that portals transcend the computer desktop metaphor and are something new. But is this really true? The now "traditional" operating system based on the desktop metaphor, like Windows, is just a tool for organizing information and functionality for its users. It started as an "organizer" for a singular computer. But as computers become connected to the Web, these barriers fade away, and the information pool that has to be administered is potentially unlimited. So, you could easily come to the conclusion that portals are just the reinvention of the computer desktop, a desktop on top of the desktop as it were.
Portals share many characteristics of the desktop, and they are indeed in danger of replicating the desktop's functionality. As the desktop is more deeply embedded within the operating system, it has more to offer in terms of functionality - portals are bound to lose if they fight the battle on this field. The desktop's problem is that it is just a set of tools, tools for organizing files, tools for connecting to the Internet, tools for running applications, etc. It is generic, maybe too generic. It is left to the users to structure and administer their electronic workplaces, and thus, how efficiently they perform their tasks or process information. The advantage of computer networks and the Web is that they do not limit users to one computer (as does the desktop), but their disadvantage is that they lack structure and user support. Portals, on the other hand, provide structure and personal guidance. This is where they shine and transcend the desktop. They offer information and functionality that is relevant to users as individuals and - as with corporate portals - with respect to their companies' needs.
Portals need a "sovereign posture," as Alan Cooper put it: they claim the whole desktop and cut their users off from the rest of what is happening on the computer desktop. The portal and the desktop are not really integrated. But should they be? Maybe it is time to reconsider whether we are doing the right thing with our computers. Are all the many programs which cohabitate on our computers necessary for fulfilling our tasks? Most of them are not. The profusion of smart utilities we install on our hard disks are only needed for solving problems with the operating system or with certain programs - they are remedies, not means. For some people, fiddling with the computer may be a goal in itself, but for the majority of people who use computers it is not. I believe that new design solutions are needed which make the distinction between the desktop and portals obsolete. Perhaps the solution is a document-centric user interface where users only work with documents but never have to bother with programs? In such an approach documents or parts of documents, for example tables, charts or graphics, "know" which program can handle them and automatically call the one needed. The users may not even be aware of this. Apple's OpenDoc was a step in this direction, but it never came to market, while Microsoft's OLE (object linking and embedding) is just a faint reminiscence of this.
With the advent of portals many authors believe we have a solution to the limitations of the desktop and the chaos of the Web. But the big challenge is still before us. We still need an effective way to organize and regulate access to huge amounts of information, and then pair that with a useful set of tools. Innovative design solutions are needed to handle this complexity and hide it from the users so that they can focus on their tasks. Portals are just the first step in this direction.