ANALYZING THE CURRENT
TASKS of a person using desktop dictation products is one
way to decide what a product should do. Such a limited approach,
satisfies the base needs of the user and gets the product out
the door. However, this approach leaves unanswered the question:
what will a user want to do with the system once he gets it? Our
testers carefully consider this last question when evaluating
a product and rate the product usability accordingly. Why? Products
which elegantly anticipate user's needs sell better because they
satisfy better.
We believe there are 16 typical ways a user may want to expand
any feature of a product. These ways group into three major categories:
- expanding the number of objects to which a feature is
applied
- integrating with other system facilities
- modifying the way the product carries out instructions
Below are the categories with their respective subparts. Illustrative
examples are also given.
1) Expanding the number of objects
- using more than one instance of a feature - e.g. dictating
into more than one program at a time alternating back and forth
between them all
- reusing a facility - e.g. being able to cut and paste
a section of a voice macro into a new macro
- inclusion in a larger goal - e.g. using voice macros to
edit a database
2.) Integrating with other system facilities
- interleaving with other goals - e.g. the voice macros which
are recognized change as appropriate to whichever window has
focus
- taking advantage of other goals - e.g. a set of voice
macros which determines that optional facilities are currently
available because a special set of voice macros has been
loaded by a previous context
- stop or postpone - e.g. stop a transcription from a
mobile recorder that is clearly going awry or postpone
completion of a transcription in order to switch another
task
- get result - e.g. get a result from one voice macro
in order to determine the next activity to be done
- external activation - e.g. allow another program to activate
speech recognition in order to get text characters for its own
use
3. Modifying the way the product carries out instructions
- progress monitoring - e.g. user may wish to know or
change whether the recognizer is treating the current
utterance as a command, a macro, text, noise, or is
idling and if the recognizer changes its mind about
that evaluation
- result detection - e.g. the user may wish to optionally
know whether the last series of rapidly uttered
commands were actually heard as intended
- rolling back to a previous state - e.g. the user may
wish to undo the window context switch implied by a
voice macro (to any depth)
- recording and retrieving - e.g. the user may wish
to retrieve the last few commands for use in a new voice
macro thus also implying some recording mechanism for
those commands and a companion searching mechanism for
those selected commands
- modifying outcomes - e.g. a user might wish to modify
an otherwise successful series of previous voice commands
in order to get a new result
- modifying for multiple similar outcomes - e.g. a way
to target multiple contexts with the same voice macro
each context being differentiated by one or more additional
phrases which select the context; a companion enumeration
facility to display various instances of the more general
macro
The examples above given for each of the general expandability
principles are not contained in entirety in any current product
although some are features contained in some products. They are
in no way whatsoever the full range of reasonable expandability
possibilities. Nor should the examples above be taken as our complete
prescription for the perfect dictation product. They are examples
given to illustrate how a tester (or designer, for that matter)
may use the principles to anticipate user attempts to search for
increased productivity.
It's crucial to distinguish between simply adding product features
and our goal. Casual adding of new product features often results
in bloatware. Our expandability analysis is designed to increase
the utility within the work environment of existing product features
by careful polishing of those features in limited ways which align
with the above principles.
If you have speech applications under development and wish to
have an analysis performed along these lines please contact
us.
More on bug testing...
More on usability...
Reference: Goal
Composition by Jakob Nielsen