Comparative Evaluation

It is clear from the examples above (Sections 2.4.1 - 2.4.4) that each of these interface types are suitable for systems for improving understanding of natural resources. The following sections will evaluate the pros and cons of the interface types by review of research that compares interface types. This review will identify the most suitable form of interactive interface for a system to instil an understanding of the dynamics of natural resources for the general public.

‘People have developed sophisticated skills for sensing and manipulating our physical environments. However, most of these skills are not employed by traditional GUI

(Graphical User Interface)’ (‘Tangible Media | MIT Media Lab,’ 2012). The MIT Media Lab

specialise in developing applications for human computer interaction using new technology.

Augmented Reality is a very versatile application platform, however it is limiting as it is mainly used by a single user, with further inherent restrictions on interactions because of occlusion problems cause by hands, or by orientation of the AR Marker (or tracked symbol) to the camera. However, projected AR, where the display is projected over a surface as in the Illuminating Clay (Piper et al., 2002) example, offers the most potential AR for our purposes, as it is collaborative, engaging and interactive, with minimal occlusion issues. Furthermore, projected AR may use touch and multi-touch as the interface. Finally, this style of AR can be enhanced by using tangible objects into a Tangible Augmented Reality system, as observed in the TanGeoMS system (Section 2.4.4).

A study by Xie & Antle (2008) comparing a virtual jigsaw, operated by a traditional mouse (VR) and menu system and a virtual jigsaw controlled by tangible jigsaw pieces (TUI) (Figure 8), observed that children found the TUI to be more engaging and playful than using the mouse (VR). The research clearly showed that children preferred the tangible jigsaw (TUI) to a virtual jigsaw (VR). The children noted that they were unfamiliar with the controls for the virtual jigsaw, whereas they could easily understand and use the controls for the tangible jigsaw puzzle because they were physical jigsaw

pieces and they behaved exactly the same as any standard jigsaw. The children also liked that they could work collaboratively using the tangible jigsaw, however they were only able to use one piece at a time with the virtual jigsaw. Xie & Antle claim that the problem with the virtual jigsaw was a combination of ‘single user access and the difficulties imposed by using an indirect interaction mode constrained to a 2D space’ (Xie & Antle 2008, p. 197).

Figure 8: Tangible and virtual jigsaw puzzle (Xie & Antle, 2008).

Similar preferences were found for adults when using the geology tangible user interface, GeoTUI. Geophysists clearly determined that tangible interactions performed quicker and more effectively than using a mouse and keyboard (Couture et al., 2008). The Geophysists benefited from the tactile response and direct control available by using their hands in a tangible user interface.

A comparison study by (Triona & Klahr, 2003) of training, using virtual and physical materials, found there was no significant difference of physical objects over virtual. In their example the results are qualified by the acknowledgement that ‘the physical interaction with the materials was unnecessary in this learning context, therefore the lack of preference may well reflect indifference on behalf of the user (Triona & Klahr, 2003). Triona and Klahr believed that the users thought the use of physical objects was gratuitous rather than a required need. This outcome implies that tangible interface designs should avoid gratuitous use of physical interface objects as they offer no benefit, and may have a detrimental effect.

A recent study on primary children, evaluating a tangible programming system called Tern to an equivalent graphical mouse driven interface, found that Tern ‘offered advantages for learning’ in two of the three scenario studies (Scharf, Winkler, & Herczeg, 2008). The qualitative observation results showed that the tangible interface seemed

highly suited for collaboration with group activities and discussions, in both museums and the classroom. Furthermore, in the museum, more children (an equal number of boys and girls) attempted the programming with the tangible system than with the computer based graphical system. In addition, the observations showed that the children switched between using the GUI and then TUI, with very few only using one system. The recommendation of the research was that a hybrid tangible/graphical interface may be the best solution for the Tern study scenarios.

Following on from this recommendation, a hybrid prototype was developed and evaluated in a live situation, where it was observed that the children preferred the tangible interface for building the initial system. However, they preferred the graphical interface for correcting and redesigning the system. Interestingly, the tangible system attracted an equal number of boys and girls, whereas the GUI was dominated by boys. Early empirical studies comparing multi-touch surface and tangible multi-touch surface found no discernable difference between the two (HP 2012). Terrenghi et al argued that tangible multi-touch systems could be replaced by multi-touch systems, as the tangible object can be replaced by a virtual object assigned the same functionally as that of the object. This argument is a functional perspective only, it does not take into account the level of engagement, ease of use, or user preference. Use of virtual objects eliminates haptic feedback and the physical third dimension beyond the screen is made obvious by tangible objects. Although, gesture and touch have some degree of haptic interaction, it is at the lowest possible level, as it only affects the finger tips when using the flat surface. People like to work with their hands, so removing the tangible objects reduces the potential engagement of the interface.

Recent experiments show tangible surface touch systems to be quicker and more accurate to use than multi-touch surface gesture controlled systems (Lucchi et al., 2010). Lucchi et al’s study required 40 students to duplicate the shelf layout of a set of warehouses, using both a multi-touch surface and a tangible multi-touch surface. The shelf units are a modular design meaning that all the different shelf configurations were made from combining units of one size of shelf. In the multi-touch system the shelf units were represented as a rectangular graphic, which is positioned by touch and orientated by gesture. In the tangible system the shelf unit is represented by a physical tangible object made to scale. The experiment measured the speed and accuracy to build shelf units to match a series of different warehouse layouts. The results showed that the

participants felt more stressed using the touch interface and that they believed they were more successful with the tangible system, even when the results showed they performed equally well in the touch system. The completion times using the touch system varied significantly, while the completion times using the tangible touch surface were consistent as well as being faster and more accurate.

The research by Lucchi et al., (2010) is one of the few examples that targets adults. Many of the experiments for learning with TUI are specifically designed for children. One reason that contributes to the success of the research by Lucchi et al., (2010) is that it targets a clearly defined problem, a niche topic - a strategy recommended by Ishii (1997). Lucchi et al’s research is validated by comparable research with results showing TUI to be more efficient than touch systems (Schneider et al. 2010; Triona & Klahr 2003). The niche topic of the (Schneider et al. 2010) research is specialised training. Furthermore this project has substantial commercial potential because specialised training courses have a strong, captive market thanks to certification requirements for Occupational Health and Safety (OH&S) as well as training requirements for in-house performance management. Schneider et al's (2010) example of the warehouse design training system is a positive example in a very lucrative market. Perhaps specialist training should be a focus for TUI, if they are economically feasible. Future research studies could define the characteristics of training that would be suitable for TUI. Not all training would suit a TUI, possibly because:

1. The development costs for either UI or hardware exceed the benefits, or

2. The problem space does not translate into a tangible user interface, for example,

tax law changes may not translate into a useful TUI.

Tangible multi-touch surfaces are easier to learn and more accurate to use than a mouse driven GUI, or a multi-touch surface (Tuddenham & Kirk, 2010). Tuddenham & Kirk conclude tangibles to be better than other inputs which validates the seminal experiments by both Buxton and Fitzmaurice (Fitzmaurice, Ishii, & Buxton, 1995).