资料来源 : Free On-Line Dictionary of Computing
thin client
A simple {client} program or hardware device
which relies on most of the function of the system being in
the {server}.
{Gopher} clients, for example, are very thin; they are
{stateless} and are not required to know how to interpret and
display objects much more complex than menus and plain text.
Gopher servers, on the other hand, can search {databases} and
provide {gateways} to other services.
By the mid-1990s, the model of decentralised computing where
each user has his own full-featured and independent
{microcomputer}, seemed to have displaced a centralised model
in which multiple users use thin clients (e.g. {dumb
terminals}) to work on a shared {minicomputer} or {mainframe}
server. Networked {personal computers} typically operate as
"fat clients", often providing everything except some file
storage and printing locally.
By 1996, reintroduction of thin clients is being proposed,
especially for {LAN}-type environments (see the {cycle of
reincarnation}). The main expected benefit of this is ease of
maintenance: with fat clients, especially those suffering from
the poor networking support of {Microsoft} {operating
systems}, installing a new application for everyone is likely
to mean having to physically go to every user's workstation to
install the application, or having to modify client-side
configuration options; whereas with thin clients the
maintenance tasks are centralised on the server and so need
only be done once.
Also, by virtue of their simplicity, thin clients generally
have fewer hardware demands, and are less open to being
screwed up by ambitious {lusers}.
Never one to miss a bandwagon, Microsoft bought up {Insignia
Solutions, Inc.}'s "{NTRIGUE}" Windows remote-access product
and combined it with {Windows NT} version 4 to allow thin
clients (either hardware or software) to communicate with
applications running under on a server machine under {Windows
Terminal Server} in the same way as {X} had done for {Unix}
decades before.
(1999-02-01)