Learned Lesson #1
The study of designers in D5.3 has provided more evidence for a definite and growing business case for adopting inclusive design, but has shown that there is work needed to make that case clear to developers of digital products and to provide the necessary support, such as VUMs, for controlling the cost and time requirements. We have shown that some major factors, including the possibility of increased market size and the need to meet legal or regulatory requirements, are likely to greatly increase in the future, but that there is still a lack of comprehensive design guidelines, standards, knowledge and expertise, as well as access to design tools and users.
Learned Lesson #2
VUMs have the potential to eradicate design errors earlier and more often in the design process than real user testing, as there is no requirement for a physical prototype with which to test. VUMs may also reduce the time required for user involvement if they can simplify and speed up the testing processes themselves.
Learned Lesson #3
Inclusive design is not defined by user testing alone. Rather inclusive design is a design approach, consisting of many different tools and techniques, including but not limited to user testing, which lead to products that are usable by and accessible to a more diverse range of users. If a VUM toolset is seen not just as a cost saving tool for industry but also as an educational tool, it has considerable potential to pass on inclusive design knowledge to designers through its use. Virtual user testing is already happening whether inclusive design advocates like it or not. How realistic those virtual users are, however, and how representative they are of real users, will depend on the success of initiatives which simulate diversity and impairment.
Learned Lesson #4
Although the VICON toolset has been shown to integrate smoothly into the design workflow and is not cumbersome for designers in their work, it has become clear that VUMs should be provided in different ways to best meet the needs, preferences and environments of different designers. For example, some designers would like the VUM integrated into a CAD environment, but others, who don’t work with CAD tools, want it integrated within their usual work environment or available as a standalone tool. Some find the 3D virtual usage simulations useful whereas others prefer a more data-oriented ‘scientific’ visualisation. These and other findings show that VUMs should not be developed with a one-size-fits-all ethos.
Learned Lesson #5
More work is also needed on creating adaptive API’s that can successfully represent many fine granular levels of the user, as well as algorithms for analysis and recommendations. Work done within the VUMS cluster towards the standardisation of virtual user models and results fed into international standardisation groups like ISO and W3C will help to further the development of standards in this area.
Learned Lesson #6
Some of the designers’ comments provide evidence that the inclusive design support proposed by VICON is actually able to help designers in addressing particular inclusive design challenges and the prototypic implementation provides additional knowledge to designers who are not familiar with inclusive design.
Learned Lesson #7
The majority of the designers confirm our belief that the virtual user concept of VICON is capable of product development acceleration. However the risk of oversimplifying the reality should be considered. We believe that the VICON system is capable of assisting designers to avoid faults in inclusive design related to particular disabilities (i.e. in this case hearing, vision and manual dexterity). However, due to the complexity of interaction between a person and a product (relating to cognitive, sensory and physical end user attributes), it cannot be denied that product evaluation with real users is still necessary.
Learned Lesson #9
An additional “task execution method” model is necessary for the implementation of possibilities how to perform tasks. This additional model based on user trials with products is necessary for the specification of parameters. This model has to be added to the inference approach of the VUM.