Earlier quoted context omitted.
To some extent we can support them ourselves (as I do with my Canon). It can can be easily set with an IP/hostname and "just works". When it's not in use, it's unplugged. The main support issues are probably physical, right? Newer printers tend to break too easily. And yes, some academics are too dumb to fix a paper jam or change a toner cartridge. Hasn't it always been so?
To a certain extent, you are correct that the physical support like jams and toner replacement was typically handled by the lab. However it still meant having 30 printers, all of them of different makes and models and firmware versions on the network. Thirty ethernet runs having to be patched in, and 30 more ports on the switch lit up, and 30 more static IP leases doled out. All of that is relatively minor on its own…
Why can't we share a communal printer? Good question. Especially considering how sparse usage is compared to the 80s/90s.
My guess is a "tragedy of the commons" type snafu, with the printer always seeming broken by someone else's job.
I actually liked going to the print room - kinda like the water cooler where you could meet people and talk - catch a look at their print job and strike up a conversation "Oh I see you're running regression tests...". IIRC at UCL CS the print room was also where the coffee pot lived.