In Defense of Toasters
Posted by Bill Venners on January 12, 2000 at 11:18 PM
> First, a very cool, very thought provoking article.
> Second, a toaster is a stupid cyberspace object example.
> One my be able to turn the toaster on or off in cyberspace,
> but the bread has to get in the toaster somehow. I think,
> maybe, in real world space one has to go to the toaster,
> put bread in it, and turn it on there.
> Lets find a more exciting and plausible cyberspace example.
Two comments. I figure that by "stupid example" you probably
mean "contrived example." If so, I agree -- to an extent.
When I write articles about programming I almost always use
examples that are contrived to some degree, often to a great
degree. (By contrived, I mean an example that wouldn't really
exist in the real world.) The reason I and other authors use
contrived examples is that they allow you to create an example
that highlights clearly what you are trying to demonstrate.
In real world examples, what you are trying to demonstrate can
easily get lost in lots and lots of real world details.
Part of my choice in using a toaster is just that: it is an
extremely simple device that doesn't have a lot of details. I
can show that a single slider control on the front of a toaster
doesn't translate well to web pages.
But on the other hand, the toaster isn't completely contrived.
Imagine I like my toast very dark, but my wife likes her toast
barely singed. One service a Jini enabled toaster could offer
to the world is it could enable Jini clients to set the
toaster slider remotely. Why? Well, even though I have to be
in front of the toaster to plop some bread in the top, it might
be nice if the toaster could recognize that its me and set
the setting to dark. When my wife is standing in front, it would
be nice if it could set it automatically to light. Some service
that is following me around my house could grab the toaster
service object when it sees me making toast, and set the
A coffee maker could offer its functionality over the network
as well. If I am able to pour in more than one day's supply of
coffee in an air tight container fixed to the coffee machine,
my alarm clock could grab the coffee maker's service object
and coordinate with it such that my coffee starts percolating
the first time my alarm goes off. When my alarm goes off again
10 minutes later, I'll be able to smell the coffee and wake
For the least contrived example, think printers and storage
devices. These two kinds of services already have working
groups trying to figure out the well-known interfaces.