Original Publication Date: 1993-Mar-01
Included in the Prior Art Database: 2005-Mar-18
Daggett, G: AUTHOR [+5]
Disclosed is a method of creating an agent to interact between a speech recognition system and an application running under a windowing system.
a method of creating
an agent to
between a speech recognition system and an application running
under a windowing system.
applications are run under a windowing system and are
driven by mouse movements and keystrokes. Current state-of-the-art
windowed applications are not able to support the recognition of
speech to drive their user interface. Although input is well
architected for mouse and key- strokes, each new input technology
must be integrated into the kernel of the windowing system, such
as Presentation Manager under OS/2*, or X-Windows under Unix**.
describes a solution
for speech input through
the creation of an agent to interact between the speech
recognition system and the application running under a windowing
system. This agent translates recognized speech into mouse movements
and keystrokes that are understood by the target application. An
application specific profile is required for this translation.
speech recognition is not an explicit selection process
like the keyboard, it is necessary for the agent to buffer the
recognized text for correction and modification. After the user is
satisfied with the buffered text, the agent uses the application
specific profile to convert the text into keystrokes and mouse
movements representing text and commands to the application. This
resulting stream then must be piped sequentially to the windowing
system for delivery to the application to drive the user interface in
the expected manner.
Window is an agent that is able to utilize the
Speech Recognition System Application Programming Interface (API) to
acquire recognized text and translate the text into mouse movements