• Hardware
• Software
• Applications
• • Healthcare
  • Emergency response
  • Maintenance
  • Production
  • • Objectives of the user tests
    • Participants in the tests
    • Task scenario
    • Experiment setup
    • Testing methodology and procedure
    • Data collection and analysis
    • Results

In summary the main findings during the first experiment were:

• Users improved their performance when using the wearable system with implicit interaction: The assembly tasks were performed faster and with less error. In fact, it took 67 seconds less in average than when paper-based information was used which was actually the second best alternative.
• Users did not learn faster using a wearable system. In fact, people were able to learn faster through paper-based support. Although the difference was neglectable when compared to those using context based interaction.
• In the test performed the day after, paper-based learning performed the best, while context-based interaction performed the worst. 
• Voice recognition-based interaction was the preferred interaction modality by workers.
• Workers preferred graphical information to text.
• Workers found the system very useful when doing a complex task, allowing hands free access to information, avoiding dispensable movements in order to check information

In the second experiment the best performance was obtained when the information was presented on the large display, and the worst when accessing through the binocular HMD. However, when asked about the user’s preferences, most of users chose the binocular HMD as the best choice.

The prototype including all features and built upon the components develop inside the project was presented during the 3rd project review at ETH Zurich, October 2006.
There has not been a formal evaluation procedure with workers at Vrachlabi, but some findings during the development of this prototype can be summarized as follows:

• The WUI, allows a hardware independent interface development, but some features should be added in order to allow more ‘usable’ interfaces. 
• The JContextAPI provides a very easy to implement mechanism to handle context information
• The prototype will benefit enormously from the existence of an authoring tool to create the procedures that constitute a training session
• In the same sense, the wearlet concept that is being developed in the framework will contribute to a more general development procedure in those cases where a procedure behind the main logic is.
• The QBIC, although having a very well accepted form factor, is not mature enough to be considered a valid development platform. Hopefully, the OQO performed according to the requirements and not extra effort was needed
• It is complicate to think in a near future for this prototype where all sensors can be removed from the car. Nevertheless the use of the prototype for training purposes makes this aspect no relevant, although it should be take into consideration in case of trying to deploy a similar system into the assembly line.
• The wristband was robust enough for laboratory use, but it needs a further development to be usable in real working conditions. There is a new design in course, that will tackle with this issue.

Wizard of OZ is definitely a very powerful tool to test some concepts even before some of the technologies are available