CHI 96 Workshop: HCI and the Web, Position Papers
Enter the Web. Individuals are now thinking of their computers as a communications tool, a tool for mining information from sources around the world, and in some cases, entertainment. The "personal" computer is now beginning to fulfill some of its original promise: its now beginning to become much more a part of the fabric of our daily lives, both at work, and at home.
HCI and the Web are then quite compatible. The Web provides HCI access to a wealth of information about users' usage patterns, perceptions of usable sites, and effective ways to organize information. In addition, the upcoming technologies which provide for more dynamic pages and user interaction provide designers with increasing tools to improve/enhance the user's experience. At the same time, as these new tools begin to provide richer interaction the experiences and techniques of HCI become increasingly important.
HCI professionals are rapidly recognizing the opportunities and synergies the Web provides for improving the interaction between people and information systems. The Web is a little slower to recognize the role of the traditional HCI on the Web. How does HCI as a discipline become a part of the evolving Web technologies? First, HCI professionals can utilize the Web and the evolving Internet technologies to help investigate and develop interfaces which enhance the user's experience. Second, as Web interfaces continue to evolve from relatively static presentations of linked pages to more dynamic, interactive interfaces to information, the role of good HCI design becomes increasingly important, and HCI professionals must be ready to play their role in the design of good Web interfaces.
On the other hand, the Web is a very different place. Often driven more by what's "cool" than by what's useful. Web interfaces need to attract users' attention like marquis outside a movie theater. And like advertisements, Web interfaces need to be new and different -- changing occasionally to maintain interest and attract users to come back often. In general the goal of a Web site is to provide useful information so that people will visit it often to access that information. Sometimes that information can be work-related and at other times it can be for pure personal information or entertainment.
I believe that the first step to excellence in design for the Web is use of the Web. There is information for HCI professionals on the Web, starting with the HCI section in the World-Wide Web Virtual Library. The CHI conferences have begun putting conference information and conference proceedings on the Web. And many HCI-related academic and industrial groups have begun making information available via the Web. Using the Web is also a great way to find usable Web sites and to experience unusable ones.
The next step, in my view, is for HCI professionals to begin using the Web to make information available about their own projects available. As they become more familiar with the technology, newer technologies provide the opportunity to use the Web for prototyping, particularly static "low-resolution" prototypes. This experience will be invaluable in working with Web developers to understand the limitations of the technology.
During the course of the year, it became clear that we were not meeting our objectives for the project. Users were not using the site on an ongoing basis. This created concerns on the part of management who wanted to use this mechanism as their primary means of providing information in a timely fashion to employees. Interviews with employees and analysis of the top 50 pages told a convincing story: employees were not really following the hierarchy, and the information on the Home Page was not timely enough. In late 1995 a project was begun to redesign the initial page users connected to. The project team included a user representative, a Web implementer, a graphics designer/HCI specialize, a systems architect, and a representative from the operations staff.
The redesign was based on two factors: access to relevant (i.e. interesting) information, and dynamic update. We wanted users to connect to the new "Front Page" (as we began calling it) in the morning and keep it up all day. In this way, we could be sure they would see important announcements and information. In order to induce users to connect to the Front Page in the first place, we needed to provide links to the information they wanted to see. We assumed that this information was in two forms: first, timely information that was relevant to the users such as press releases, company announcements, etc.; and second, information currently within the Web that they visited often as measured by our access logs.
The Front Page uses a newspaper metaphor. After an identifying banner (featuring the Wizard character used throughout the site), there is a date, and a possible headline. The headline replaces the previous "HotFlash" mechanism without sacrificing as much screen space. Also on the Front Page are links to the most popular pages, a subset of the announcements posted that day, and a graph of Genentech's stock for the day. In addition is a "Feature", which varies from day to day. The feature is used for a number of things, including the regular Friday feature: Net Roulette. The page is implemented using Client-pull, and refreshes itself every 15 minutes. This way, dynamic information (headline, announcements, and the stock graph) change every fifteen minutes, keeping some interest. The feature can also change on a periodic basis to provide increased dynamics.
The response to the Front Page was extremely positive. Users are keeping it up more often, and the total number of hits in the system has increased. We are already considering enhancements based on using frames and possibly a Java stock-ticker for headlines and announcements. The latter in particular is in response to requests by information providers to provide increased visibility to particularly time sensitive "hot" items such as computer or network outages, facilities events, or critical company announcements.
After several months of work, the site was shown to management, who were enthusiastic in their response, but wanted additional review of some of the site's content to ensure that we were well within the US Government's Food and Drug Administration guidelines for advertising. The result was that the site was never released to the public, even though it provided useful information, was very usable, well designed and attractive, because the process for release was never clearly defined, nor were the requirements for the site. Genentech, as a pharmaceutical company, must adhere to the guidelines set forth by the US FDA. No representative from Regulatory Affairs was a member of the team. In addition, at a key point in the project, the management of the overall effort was handed off, resulting in a loss of focus. Since there was no regular project team meetings, the project languished.
In retrospect there are some important lessons that can be derived from this experience. The development of a significant Web site is a systems development project. A project team needs to be assembled with representation for interested parties, developers, designers, and management. A smaller team should be assembled which will actually be responsible for the design and implementation. The larger team provides input and feedback on the evolving design, as well as providing content review for potential management or regulatory issues. In the case of Genentech's external Web project, there was also no clear HCI responsibility. The graphic designer was not even an official part of the team, and no other representative on the team had an HCI background. Hopefully, in the near future, we will be able to revive this project along these lines to determine if the difficulties were in project structure, the newness of the technology, or regulatory uncertainties associated with presentation of information in a new medium.
CHI 96 Workshop: HCI and the Web, Position Papers