Quoting Hendrik Boom (hendrik@???):

> Round about 1990 I was using an X terminal. 8 megabytes of memory,
> impleented the X protocol, and almost nothing else. It presented a
> login screen on sshich I could tell it which coputer on the network I
> wanated to log in to, as sell as the usual name and password, and after
> that I had X with a window manager. If I wanter a so-called desktop,
> I'd tell the remote machine to start it.
>
> It worked just fine. Nothing special needed on the remote machine.
>
> THe X terminal could even boot over the net. It did not need much in
> the way of local permanent storage.

Yes, indeed. I've done many of the same sorts of things -- and options
like X2Go remain extremely practical and well-maintained, providing high
performance and first-class functionality.

This (your 1990 X11 terminal, X2Go, etc.) is all _network-based_
multiuser, of course, using intelligent remote workstations. The aim of
multiseat is to implement _locally attached_ multiuser, on attached
terminals -- which is different in kind (IMO an almost entirely
pointless niche use-case, but proponents disagree).

I like how Arnt phrased it: 'Multiseat is unimportant, barely
significant. The price of computers has dropped enough that the ones
with UIs are now personal devices.' The economics if the multiseat
aspiration doesn't, IMO, make sense in the 2010s. People who disagree
are of course welcome to scratch their own itches, though it'd be nice
if they didn't drag the rest of us along with them.