I think I have lost the part of the method we are currently arguing about at the moment. Could someone clarify?

Presumably, if there is a Spider monitoring the node or Active IC with scripting that covers the situation. This is either up to the PC's to actually have or the GM to decide when crafting the systems security profile.

If he's just winging it and hasn't concluded that the system needs a security profile then the answer is probably no, nobody notices...

Say the spider/PC is not spending IPs on "Overwatch", would they get a Matrix Perception test to notice the roll? Or is this only for Active Monitoring?

Wait. I am going to read the section of Unwired. This has been enlightening. I just hope we don't get into scripting at the gaming table

BlueMax

Think of it this way... you program agents w/ simple scripts, then they use their primitive fuzzy logic to follow the order.

Simple examples are like above... use browse and search for this data for me.

More complicated examples.
"Scan every Icon which enters the System"
"If you detect an intruder:
send me a priority alert
activate the combat ice (another agent)
Move the access logs to the data dump, replace w/ dummy log, and deactivate it's wireless).
Start Tracing intruder, message me when you find his location"
(EG: a piece of cyberware slaved to the commlink, dump the file there, then deactivate it's wireless so it'll need to be reactivated via DNI).


Or for example... an agent programmed to patrol your allies slaved commlinks and make sure they're all set to the proper settings (broadcasting safe SIN, not running hidden in a high sec zone, etc.). Or an agent which runs your EW... regularly changing your encryption or accessID. An agent which does nothing except sniff and snoop on others transmissions. Even makes short decryption attempts on the theory that at least a handfull will be decrypted, and some might be interesting.

It's almost always best if you write out a simple program sequence for your GM (or if you're the GM write out a simple program sequence so you can quickly resolve how the agent reacts to it's scripting. If it's not clear roll it's rating in the same way you'd roll a drones pilot rating if it encounters an unforeseen obstacle.