David,
Dumb terminals (e.g. 3270's) have a screen buffer, a keyboard
buffer, and firmware that allows them to connect and sense
the transmission media (e.g. coax-connected 3274's). When you
turn them on, they have the ability to send an "I'm here" message
to a communications controller. The rest of the embedded logic
comes directly from the keyboard which can trigger a basic send
and receive sequence and type characters to the screen and the
screen buffer which actually does the sending and receiving of data. There is minimal error reporting, mostly relating to hardware
failure. That's it.
The NC has a remote bootstrap sequence that allows you to fill
memory (32MB or more) with an entire OS or a Java VM. The NC
is really an intelligent terminal, not a dumb terminal. The
applications are loaded from a server but a number of software
based optimizations such as cacheing and intelligent application-
level communication protocols limit excessive sending and
receiving of redundant (persistant) data, thus making the NC
more efficient.
Bill Gates is too stupid to design and develop an NC, & he looks
at anything that isn't a fat client with contempt anyway, so
his response to the NC is: "OK, the market wants cheap workstations,
we'll give 'em cheap workstations, we'll sell dumb terminals and
call them NC's. Who's gonna know the difference anyway? The
public is too illiterate to comprehend all this nonsense Sun keeps
talking about. It's all in how we market it. We'll sell a bunch
of them, they won't work too good, everyone will say that the
concept is a bad one and the techonology is old & outmoded & they'll
go back to their PC's."
cheers,
cherylw |
