NUI, or better yet, NLUI (Natural Language User Interfaces) are
starting to appear at a high rate in directional phone systems that
prompt the user to specify the "natural" answer to a series of
questions. The system relies on voice recognition and it is the first
practical application that is being provided to the public. However,
it has it's limitations. It generally takes longer for an able person
to listen to the prompt and answer clearly than it does for them to
press a number on the phone board. It doesn't always understand what
the person is trying to say due to line noise, etc. This is
expecially true with cellular clarity issues, and grows daily with the
increased usage of cell phones.
A GUI is much more practical for most applications today where a
human/computer interface is required. This is because it performs
translations so that inputs by the human are understood exactly by the
computer system. Natural language is often ambiguous due to so many
factors it's not practical to cover all of them here. But most
notably, vocal fluctuations, communication patterns that differ from
culture to culture, etc.
NLUI's have their uses where they excell beyond the abilities of a
Graphical User Interface, but these are typically most benificial in
areas where people are not capable of harnessing the benifits of
learning to use a human to computer interface such as a GUI.
Handicapped persons, for example, may be unable to see a keyboard or a
screen, and therefore a natural language voice recognition system
might benifit them. But also consider that there are many ways to say
the same thing in any language. So there is still some "unnatural"
aspects to most NLUI's... Dragon Speach had (and may still have) a
software program that allowed the user to narrate text input to the
computer. While this greatly assisted people who couldn't type by
hand well, or at all... It limited what the user was able to do in
terms of telling the system how to respond to editing commands because
they were in the realm of human to computer communication whereas the
computer could care less about what you were saying to it when it was
simply converting your speach to text.
I think NLUI's are a ways out still. The processing power required to
interpret what a human being intends to do at the rate of speed that
we are capable of operating at with GUI's is astronomical, and
requires the assistance of AI (artifial intelligence) which in it's
prime will be enormously resource-intensive. |
