next up previous
Next: SUMMARY Up: SPEECHWEAR: A mobile speech Previous: FIELD TRIAL

GENERAL OBSERVATIONS

The development of the SPEECHWEAR system was a success: a working system was produced and was tested in the field under conditions of actual use. At the same time an extensible infrastructure was created (SPEECHWARE) that can be applied to a variety of domains based on hypertext multi-media documents.

The experience also revealed a number of problems with this approach. For example, the form of the task as designed followed quite closely that used in the original VuMan implementation and was implicitly constrained by the characteristics of the rotary mouse interface. Analysis of the task structure, for example, suggests that a different protocol (implicit confirmation [8]) could eliminate approximately half the steps in the original task, by implicitly channeling the dialog along the most likely path and relying on the user to indicate deviations. An analysis of the data showed that about 90% of items were judged Serviceable, yet the protocol required the user to input this item explicitly, then confirm it. A simple confirmation of a suggested default input (Serviceable) would have been sufficient to enter the inspection outcome.



Alex Rudnicky
Thu May 30 19:32:28 EDT 1996