Frankly, I'm surprised you wouldn't know this after being in the industry for 18 years. It's the sort of thing that becomes quite obvious if you have any experience with operating systems.
Not necessarily. At an old job, we had a small "server" (an HP PA-RISC workstation) running a small Oracle database. That machine at one time had not been rebooted for three years, with no issues with it being up so long. And noone noticed that it had been up that long.
This day and age, hopefully, most large IT organizations have a regular maintenance schedule, which handles the reboot after applying new patches. Add in redundant systems and systems being more robust/hardy at the OS level these days, probably less need to reboot. Why we have systems out there with 99.998% uptime. And noone really looks at it being a "just to be safe" reboot.
More likely to have issues with the service(s) running amok (eg. memory leaks) on a server. And probably can handle that by shutting down the service and restarting. But even then, becomes something like above re: regular maintenance schedule.
Now, yes, with something like an iP* device which is a consumer sandboxed environment with little options to do real admin work on them, an occasional reboot helps with poorly designed apps. Personally, I do it once a week, but probably could easily stretch that out to two weeks, once a month, as to date, have not seen any odd device behavior that needs a complete shutdown (the occasional swipe quit/close of an app is enough).
