I agree.
And this is still consistent with what I've been saying.
Not at all, you specifically stated "Although, if you have that many things running at once, I would say that the stability is definately more susceptible to hiccups.". But the truth is that the amount of processes means absolutely nothing, the only time the amount of processes will affect the system is in the very extreme case where you run out of PIDs or memory to keep track of processes and those are extremely unlikely scenarios.
You're the one that's mentioning getting a virus, when that has nothing to do with the number of processes or how that relates to stability. You're adding variables to the discussion that don't belong. And nobody said anything about disabling important services or programs.
You specifically stated that adding an antivirus doesn't add stability, they add features. But the fact of the matter is that those features generally add stability since the alternative is using an infacted machine which will almost always have stability problems in addition to the infection.
A security problem? Laffo. OK then, how about you go to the Services MMC, locate "Remote Procedure Call (RPC)", open its properties, stop it, and then set it to Disabled. And restart. Then if you have enough functionality left, try coming back here and report how well that's working out for you.
I realize it's required for alot of Windows' functionality, but that doesn't mean it's any less of a security hazard. IE has been riddled with holes in the past as well, but that doesn't mean you can actually remove MSHTML and have things still work. On unix if you want to use NFS you need to have portmap running, it's a conscience decision to allow a potentially bad service to run so that you can get some functionality that you want.
Please, prove to me that these processes somehow end up taking only a few bytes, and aren't doing any other sort of actions.
Go to your local book store and read a few chapters of Inside Windows. A sleeping process will never do anything unless activated by an outside source, if the process stays idle long enough Windows will evict it from memory to make room for things that are being used. The only thing that won't get paged out is the process descriptor since that's part of the kernel's memory and is needed incase the process does want to run again. On Linux a task_struct is 960 bytes, I can't imagine the NT one would be a whole helluva lot bigger.
Fact is, several unnecessary processes WILL take up more than just a few bytes, or even kilobytes for that matter, and they end up using other resources as well; whether it be for starting, stopping, or even while it's seemingly sitting idle. Resources such as: I/O reads/writes, processor time, threads, handles, etc. This doesn't amount to just "a handful of bytes".
Fact is, you havn't done any kernel development so you don't know any of that for sure. When the service starts, yes it will use some resources. But after it sits idle for a bit they will be reused by the NT VMM.
I/O reads/writes are a one time thing, at request of an I/O the kernel will perform the action and free the memory to be used for I/O for another process.
No CPU time will be allocated to a sleeping process. Unless the process has a reason to activate it will sit 100% idle.
A thread is a lightweight process, it takes even less memory to track one of those.
File handles are normal integers on unix, even if the NT ones are twice as big for some unknown reason they'll using a whole 8 bytes of memory.
