I suspect that what happens to some STBs/people is that the firmware (in RAM) becomes corrupt somehow (much like it does on many computers if you leave them on and never reboot). A reboot simply reloads the firmware and the issues go away, just like a reboot often cures an issue on a computer.
A monthly reboot of the STB is a recommended procedure in the "tips". (turning the STB off doesn't perform a reboot of the STB, it simply sends it into "standby".)
That's an excellent explanation and an excellent suggestion.
I recently retired from a company after spending more than 25 years debugging multi-million dollar custom written industrial control software which was purchased from contractors at the same time I was originally hired.
In the beginning, our systems (we had about a dozen of these) could run no more than 24 hours before crashing. Over the years, our team slowly sorted through the bugs and at some point, our systems were staying up for as long as 6 weeks continuously before an incident.
Sometimes there would be a power failure which would put an end to the endurance test, sometimes it was unscheduled hardware maintenance by our technicians.
But we never had any doubts that for some hardware or software reason, these systems would crash sooner or later.
Rebooting a SA8000 series PVR on a regular basis is an intelligent approach in dealing with a complex system which no doubt will be more vulnerable to having a problem within its volatile RAM memory as they age.
It would be nice if one could program into the PVR a moment in every week when a scheduled soft re-boot could be achieved without bothering the owner. Afterall, if Videotron can randomly download a new SARA version, surely a reset can be automatically triggered locally or remotely on a regular bases.
But would a company like SA be embarrassed to admit that this is a reasonable and necessary approach that goes hand and hand with ownership of a complex machine?
We have two computers in our home office and each gets rebooted every day so we have no idea how long these would survive before a crash. But in the basement is a computer running Win98 which looks after my HVAC system and it will run 26 days before the internal clock counts through its full cycle and causes an overflow error in my control program. This is a known problem with programmers who use the Visual Basic GetTicKCount function. Most of the time, Win98 stays up for the full 26 days. Occasionally a power failure causes a reboot.
I will be adding the missing protection against the clock overrun to my program and then see how long Win98 can crunch away before crashing.
How many micro-computer based appliances do you have in your house which are on all the time and what is the longest time they stay up before being corrupted?
For me, it would be my less sophisticated Honeywell programmable thermostat which has been up 15 months since I last changed the battery. It can't afford to crash, else my pipes might freeze while I'm away on my winter vacation.