As mentioned before, presenting a list withalldisplays in the surrounding may lead to increased selection times due to the limited screen real-estate present on mobile devices. Several displays can be excluded a priori if they do not meet the minimum requirements for a certain task. For example, viewing pictures stored on a personal device may not be wanted on a display built into a ticket machine due to its size and (possibly) frequent usage. On the other hand, obtaining a ticket for public transportation has to happen on a display that is actually capable of issuing
4.1 Selecting the Target Display 67
the ticket. Hence, the user needs to state these minimum requirements in the very beginning. Considering numerous properties, however, this may be cumbersome and time-consuming. On the other hand, embedding the connection process into an application, may overcome this by letting the application define the minimal set of requirements of a target display. Similarly, if these properties are defined for each target display as well, an incoming request from a mobile device can be matched to determine the suitability of the corresponding display.
Properties describing displays can be broken down into two major categories -staticanddynamic
properties. Static properties define characteristics of a display that are not easily (or not at all)
changeable. Such properties include the display’s size (either physical size or pixel resolution), or the space it resides in (e.g., privacy-related information). They can be preconfigured in a display once it is set up in the environment and most likely will not change over time. Dynamic properties, on the other hand, may change more frequently than static ones. These properties, for example, include the availability of the application binaries on a certain display. It is apparent that this property may change over time since additional binaries (i.e., applications) may be installed on or removed from the target display. In contrast to preconfigured static properties, the dynamic ones need to be determined at runtime. For the aforementioned characteristics this appears to be a trivial task. For example, the software running on a display may check for the corresponding application once a request arrives.
Static:
40 inches or more
Dynamic:
Has Binaries
Figure 4.1:Sub-selection of displays with static and dynamic properties: first,staticproper-
ties are used to decrease the amount of displays (here: screen size). Subsequently,dynamic
properties are used to determine the remaining displays and present them with theirspatial
andtemporalproperties. The faded displays have been discarded during the process.
Dynamic properties may also include the user. The most common one is thespatial availability
of a display. It describes whether the display is closer than the maximumdistancethe user would move. In large spaces, this property may be a powerful determination of a display’s suitability for a given task. For example, if users start a display discovery request while they are located on the one end of a shopping mall, the display on the other end may not be the preferred one. To determine this property, the user needs to be tracked (i.e., the environment knows where the request approximately is coming from) whereas the positioning accuracy needs to meet the environment’s size. This means that the smaller the environment is the more precise the tracking needs to be in order to reduce the resulting list of potential target displays. The closest display (i.e., the display with the best spatial availability), however, may already be in use by another person. The display may be close but cannot be used at the moment.
Certain tasks may require the use of a display for certain time period only. For example, users should be able to obtain a ticket for public transportation within a short time (assuming a good and understandable user interface). Furthermore, the usage of large displays may be restricted in time as well in order to also allow others to use the display. One concept in doing so is to grant a
leaseto a display for a certain period of time depending on the task. In such scenarios, a display
may be unavailable at the moment, but is usable again once a lease expires. In this scenario, the environment is able to determine the maximum waiting time for each display. We refer to this component as thetemporal availabilityof a display. Since this property will change frequently, it also falls into the category of dynamic properties. The combination of both spatial and temporal availability then results in a more complex, overall availability of a display. Determining the best suitable display for the user based on this complex property is a rather difficult task as it highly depends on the user’s preferences. For example, a display close-by may be in use for the next ten minutes while another one further away is immediately available. The system does not know whether the user prefers waiting over walking to the display further away. While both spatial and temporal availability are important for the overall selection, they need to be presented to the user for a final decision based on individual preferences.
As shown in figure 4.1, the static and dynamic properties can be used to limit the set of potential target displays. First, the static properties already filter out displays that are not usable in any case (even if they would be in range and currently available). Subsequently, the filtered displays are checked whether they are capable of running the application. The remaining two properties (i.e., the displays’ spatial and temporal availability) may then be presented in a two-dimensional view to allow the final decision being made by the user.