[kidsUI] Rerun

[Sanjay Chandrasekharan <schandr2@chat.carleton.ca>]

Hey folks,

Here goes my first posting to the Agent list. I am making a business case for
Agent Interfaces here, so you will not find much to do with kid UIs.

I see the kid UI for Linux as a way of breaking the hold of the Big Great
Window. To get out of the window mentality, we need to come up with a new
interface metaphor. My candidate for it is: Agents. I consider Agent Interfaces
as the first step in breaking away from the Window paradigm. Now, read
on....this is looong..
>
>
>
>
Assumption: we need Agent Interfaces
Question: why?


The interface designer who is asking "why" is looking at it from a usability
perspective. Basically, he is saying that the current interfaces are more than
adequate, as far as usability is concerned.

My reply to that is: yes, but the model-T (or for that matter, Toyota Camry)
was adequate as well. Why do we have Porsches? The point is this: if you stress
on usability, we come up with a one-size-fits-all model of business. And the
result is that we have just one version of software for all users.

I think the usability line of thought is something every new industry starts
off with. Think of cars or clothing. As the industry matured, the stress on
usability gave way, and market-making took over. So we have shirts with
different textures and cars that look like frogs! Why? Because people want
them!

Here's another good way of looking at it. In the beginning, every new industry
carves out the population in terms of usability, as populations who want the
products to do different functions. Products are tailored to do these
specialized functions. So we had buses, trucks, tractors etc and different
kinds of dresses useful for different users. This is the phase of the software
industry now. We have programs that help writers, lawyers, doctors, architects
you name it. We are perennially on the lookout for new user groups.

But once all the user groups get saturated, the population needs to be carved
out differently, this time in terms of satisfiability. In this division,
usability IS TAKEN GOR GRANTED. What you are selling is not a new
feature or a new way of using a feature. You are selling an experience, a
personality statement, a lifetsyle. I am suggesting that the software industry
should shift to this model of business. And to do this we need agent
interfaces.

Now of course, the WHY cry is louder. Why do we need agent interfaces for this?

Well, the answer is simple. A consumer market is a created market. To create a
market, you need to create an image associated with the product. How do we
create an image for a product when your interface consists of a lot of
rectangles, objects and cryptic, one line, messages?

Enter agent interfaces. An agent interface allows you to create personalities
for software products. You can have products tailored for any age-group, any
lifestyle. So the agent interface, to me, is like the packaging of a
consumer product. It is something that makes people identify with the procuct,
and buy the product.

What do I mean by an agent interface? To start with, animated characters. For
everything, right from functions to error messages to icons. As I pointed out
in my first message, the poor MS Agent is stuck in an environment that has JUST
objects. My idea for getting out of that situation is to create programs
that does everything using animated characters. So the help agent will become a
natural part of the environment.

What about the intelligence? Short answer: it can come whenever it comes! :-)

Long answer: we don't have to wait for intelligence to make agent interfaces.
The current general idea is that an agent interface is the "face" of an
"intelligent" program that does some fancy AI work under the hood. That doesn't
have to be so. People were moving around objects on desktops long before fancy
architectures like CORBA and COM came along to do such work under the hood. So
why should we wait for intelligence to introduce agents? Intelligence, as and
when it comes, will have a natural environemnt once you implement the business
model I am suggesting.

Now, this doesn't mean that I am not interested in the intelligence issues. I
definitely am. Otherwise I wouldn't be doing Cogsci! I am just trying to come
up with a solid business justification for using Agent interfaces.

I hope that makes sense!

Sanjay


-------------------------------------
Sanjay Chandrasekharan,
Ph.D Student (Cognitive Science),
Carleton University,
Ottawa, Ontario,
Canada K1S 5B6
Phone: 520-2600 ext 1568

http://wabakimi.carleton.ca/~schandr2