For a whole system this heuristic is usually reasonable, even now. We're talking about a single process, here. |
IMO it's also reasonable for a single process. BTW, also the CADAM system consisted of several 'disconnected' machines (service "user" without console attached, running in the "background") it was the
single process of the interactive "user"/machine that caused trouble and had to be forced/taken off the run-list by operator. (And - sorry - by "user" I do not mean the human one sitting in front of the screen, for me a "user" is a virtual entity consisting of virtual CPU, virtual spooling devices (puncher, reader), virtual printer, with or without real console attached.)
That said, I don't know what you mean by "end a process immediately when TaskManager reports 'no reply'". |
Wrong wording, by 'no reply' I meant "not responding". And with
immediately I mean instantaneously/at once/in no time. And 'end a process' I mean the options to do so offered by Task-Manager, 'end the task' or 'end the process' or 'end the process structure',
interactively.
The system never automatically kills non-responding processes, although it may give the user the option to kill them. |
Yes, I know, there is no "programmable operator" on the PC, it's me, on the PC I am the only operator (I hope).
To say it in your words, it
seems rather silly, honestly, to distinguish a single process from a
virtual single process.