Timetables are a familiar and, if well designed, easy-to-use way to find the times of regularly occurring events such as films, public transport and museum tours. The interaction is simple: the timetable presents the time of each event in a grid indexed horizontally and vertically by time and date. Timetables on paper work well because they enable us to browse the complete set of available information. We can see at a glance on which days a particular film is showing, the intervals between buses, or the time of the last museum tour of the day.
Timetables presented on websites often lose their power in the translation onto computer. Computerized timetables often begin not with browsing but with search. Search is a natural interface for any data set because it’s very easy to implement. However, search interfaces tend to be data-centered rather than user-centered. For example, websites that present bus timetables often focus on how the bus company thinks about buses, which is, naturally, in terms of bus routes. Passengers must select a route before being shown the timetable for the route. This model quickly breaks down when passengers are not familiar with the routes offered by the bus company.
The initial search for a route is not required at bus stops because bus stops present the timetables for the routes that pass the stop. In effect, the geographical location of the bus stop performs the initial search that’s required on the website. However, rethinking the initial website search in terms of the passenger rather than the bus company solves the problem: given the address of the origin and destination, the website can select the route that best matches the journey. Having selected the route with user-centered information, the website can then present the timetable for the route.
Booking a flight online requires a dialogue in which passengers must tease out of the booking system enough information to make a decision. If all the flights from the departure to destination are full on a particular day, passengers must form a strategy to find an available flight nearest that date. One strategy is to look at the day before. If no flights are available on that date, should passengers check two days before, or should they check one day later? Of course, passengers follow this exact strategy when browsing printed timetables, but the interaction is natural and involves running a finger across and down the rows and columns of departure times and dates, possibly using another finger to cross-reference another date.
When airline booking systems show all the flights for a date, plus flights on several days either side of the date, passengers can see what flights are scheduled and get a sense of the pattern of available flights. By further indicating which flights are fully booked and which have seats available, booking systems enable passengers to see at a glance when they can travel. This is the power of browsing over search: passengers are presented with and can take in a large amount of information—departure times and seat availability for several days before and after—almost effortlessly and without having to ask.
Although browsing is the ideal interface for timetables, the World’s flight schedules are too large and complicated to show onscreen in their entirety. Passengers need to drill down to make the amount of information manageable. However, at each stage of drilling down, the amount of information is rapidly and massively pruned. For example, starting with all the flights to every destination in the world, selecting the destination city reduces the information to just the flights that land at that destination; selecting the departure city reduces the number of flights still further.
Timetables are a good example of how rethinking the way in which users interact with data leverages computational power in a form that’s powerful for users.