Workstation
en.wikipedia.org
en.wikipedia.org
IMO precision of task-focused manipulation over a quantity of time is a really important metric justifying workstations' existence. The hardware was typically focused on reducing time-based bottlenecks. But today a lot of the time bottlenecks are found in software and hardware interface.
To me this also implies that a workstation in hardware must be accompanied by workstation-class software, which ought to be considered it's own category of software. Workstations are best operated by a workstation operator who is trained in workstation psychology and what it is best at. That is, metric-reflective structured manipulation over time.
In the FOSS world I'd consider Blender even circa 2002 as really great example of workstation-class software for example.
IMO it would be interesting to see a workstation Linux distro which provides de facto workstation-type software plus additional workstation-type tooling to reinvigorate this niche.
The concept is currently more out of focus (or "blurred" as the article puts it) than it is obsolete, IMO. I'd guess most of today's office workers need better support for workstation-style work tasks than they currently get, at least a few times a week.
To me the ultimate workstation might also involve a method for tying together discrete software packages via sets of nodes or visual pipes. Kind of like an Eagle Mode for apps, in terms of navigation if not layout anyway. Right now far too much systems-logic strain is placed on the cognitive energies of the operator, and those are too variant to be reliably interesting for business...it's unfair to both operator and system. "Next you go over to App Y after exporting from App X and hit Import, then navigate using the file picker" could be far more easily metabolized as "drag a wire from App Y's output to App X's input".
Anyway! I love this stuff...