According to page 225, up to System programs can be run without issue. After that, you subtract 1 for each System number of programs running. Thus, for you, 1-5 is no reduction, 6-9 is -1, and 10-14 is -2, and so on. This seems odd to me, really, that 5 is included in the first penalty group, but 10 is not included in the second. However, this is how the rule is written, and it is even an example.

Yes. Shutting down a running program is a simple action. Running a new program is a complex action, according to the actions table on 219. (This also applies to starting / stoping agents)

Many people have expressed interest in this, I too would like to see it.

I lived there a while myself. Hopefully I'll move back to that general area again soon, my SR group is there still Don't miss the weather though...

System is capped by response, the only "sorta" is if you upgrade response, your system can function at it's full value. So, for all practical purposes, your System rating in your example is 5.

With System of 5 (or System capped to an effective 5 by a response of 5), as I said, you can run the following number of programs with the following penalties:

1-4 programs: No Penalty, response 5.
5-9 programs: -1 penalty, response 4.
10-14 programs: -2 penalty, response 3.
15-19 programs: -3 penalty, response 2.
20-22 programs (there are only 22 programs): -4 penalty, response 1.

It can't possibly work that way, because if System is really capped by Response then when Response decreases so does System:

1-4 programs: No Penalty, response 5.
5-7 programs: -1 penalty (System*1 programs active), response 4, System 4
8 programs: -2 penalty (System*2 programs active), response 3, System 3
9 programs: -3 penalty (System*3 programs active), response 2, System 2... but now you're running more than System*4 programs, so response decreases again to 1, which decreases System to 1... which decreases response again to 0, decreasing system to 0, and your Commlink crashes.

Ah, the Cascade argument. You can't possibly think that's how it was meant to work, can you?

No I don't think it does, which is why I was saying that Response doesn't actually cap System, at least in terms of limiting you to (System) rating in programs; it only limits you when you're actually rolling your System program rating for a test. So, for Response 5 System 6:

1-5 programs: No Penalty, response 5. System tests made with System rating 5.
6-11 programs: -1 penalty, response 4. System tests made with System rating 4.
12-17 programs: -2 penalty, response 3. System tests made with System rating 3.
Etc etc.

It still does; it just isn't RE-capped every time response decreases from system load. Thus, when response is decreased by running programs, system stays the same.

At least, that's how I interpret it -- the book, unfortunately, isn't especially clear on the matter.

Good point, I hadn't thought of that.

Which brings me to an important point -- Agents seemingly let you get around the program limits. Agents count as one running program. Agents can run multiple programs themselves. Thus, you could load all your attack programs into one agent, and run that agent, counting as a single program. The only drawback is that then your utilities are limited by the agent's rating instead of your skill, and your programs are vulnerable to the agent being attacked.

Of course, if you have sufficiently advanced agents, they may actually be better than your skill, thus leading to the various discussions around here of a "hacker in a box" -- an advanced agent that replaces your team's hacker entirely.

Ugh, but that would create a functional difference between a commlink with Response 3 running System 6 software and a commlink with Response 4, lowered to Response 3 through program load, running the same System 6 software. I'm not sure that makes an ounce of sense; why would there be a difference between two commlinks that both have the same Response?

Eh, I guess I can see where you're coming from, but with your idea you'll have to record three values for System: it's actual rating, the first modified rating when the "natural" Response slows it down (for determining max programs), and then the second modified rating when the "modified" Response slows the System down again. It just screams inelegance to me, I guess.

Even assuming that your System does count as 5 for EVERYTHING while running on a commlink with responce 5 - you can still run 5 programs without penalty, since the number of programs is not greater than the system rating.

At least, that was my take after reading the rules.

I'm not sure I see how that post is relevant to the conversation (of programs bringing down response, bringing down system). Granted, I skimmed that post - because its frigging huge - but would you mind linking some quotes or response numbers or something?

If all you want is confirmation that Response 5 System 6 (capped at 5) runs up to 5 programs without penalty, scroll up a few posts -- that was the specific example I gave, with that exact result.

This was the thread that inspired me to write that other mammoth thread, and pre-dates it

I wish I knew. Unfortunately, the rules really don't say very clearly.

What I've been going with personally is System is capped by the response when no programs are running for purposes of subscriber lists and number of programs that are able to run without penalty. For these two purposes, only hardware changes effect System. For purposes of maximum program rating, which can never be higher than System, System is capped to the current response, factoring in program load.

The rules as written may not back up this interpretation; but they aren't very clear. However, as pointed out before, re-capping System with every response decrease can rapidly lead to a complete crash. In addition, random things connected to your PAN would start dropping out. While my system initially sounds more complicated, it actually simplifies things because you don't need to re-calculate the system load or subscriber lists constantly.

On the other hand, allowing System to be reduced by load for purposes of subscriber lists leaves open an easy exploit. Hack in to the enemy street sam's commlink. Begin uploading agents (no test.) Within a few initiative passes, his commlink will start to loose connections -- smartlinks, wired reflexes, etc. A few passes more, and the commlink crashes. Once access at basic user level is achieved (easy to do when hacking on the fly) no further tests are necessary. This could still happen with my system, but it takes considerably longer.

Sometimes I wonder if the people who wrote the hacking rules had any experience in computer security at all.

I don't have my book with me, but I am pretty sure that it specifically says that commlinks do not have basic user level access. They only have administrative access which is much harder to achieve.

Since virtually everybody is supposed to have a commlink, wouldn't commlinks be "standard electronic devices"? Also, a commlink is your personal communication device, what need is there for other accounts?

I don't have the exact quote on me, but IIRC, the book defines "standard electronic device" as any device that has a "rating" (as opposed to "Response Signal System Firewall" stats.)

No where in that are such devices called standard.

im not entirely sure that smartlinks and so on need to have a entry on the subscriber list. by the sound of it, only external comlinks, drones, agents and similar need a subscription...

The commlink controls your PAN. If you want your smartlink to connect to your cybereyes you would want to have them going through your commlink for security (your commlink would have a better firewall), which means that each would be a subscribed device.
Without having them connect to your commlink you would either have to let them broadcast signals to each (allowing for direct hacking) or have a skinlink to connect just those to devices (more devices connected over skinlink would require connection to the commlink to manage the network).

You're right, I was apparently mis-remembering.

The only other place in the BBB that "Standard" is mentioned with "Electronics" or anything like it, that I was able to find, is a table on 214. All it mentions there are "Standard Personal Electronics," Which, even if it where the same things as a "Standard Electronic Device" (and not, for instance, a subset) no information is given on what this term means.

I guess it's yet another term in the book that just wasn't described. If it wasn't for my feeling that, in general, account privileges are hosed, I'd be frustrated by this.

NODE-Any Device or network that can be accessed.
SR4, p.216

It depends on how things are set up. The boundries of a node are typically defined by whether or not you'd have to log in with an account and password.

If you set up your cybereyes so that you can toggle your eye-recorder on and off from your commlink without having to use your commlink to log into your eyes, then commlink and eyes can count as one node.

If there's an individual computer controlling each video camera in the building, and these computers aren't set up to communicate with each other, and are set up that someone who was in the central office would have to remotely log in to manipulate the camera controls, then each video camera counts as a seperate node.

Spoofing commands to a drone is gaining control of it, but from the thurst of your questioning, I belive you're looking for a more permanent solution. That would involve logging into the drone itself with an administrative access account to configure the drone's subscriber list. Smart riggers would require such a login to come from a hardwired connection, to prevent someone from reconfiguring the drone over wifi.

As for hacking the commlink, if you can hack yourself an admin login without the comlink's user noticing, then yes, you can issue commands to any drone. However, the complexity in being able to pull this trick off for long without the commlink's owner noticing would make this prohibitive.

Data search (p.220) is a good starting point.