Gming a game, the Matrix is still pretty fuzzy to me and my group. I'd appreciate some answers on a few questions. For ease of reference, all page numbers will be using SR4a, unless otherwise noted.

Deleting information off a commlink. How does that work? 230 seems to indicate the edit software, but that seems odd. For one, it means that deltree *.* is harder to pull off then actively changing a face on a video feed. Two, it also seems to indicate that it's something that has to be actively done over several passes. Lastly, can this information, outside of GM/CSI fiat, be brought back?

Black Ice/Hammer/Out. It prevents the hacker from jacking out unless they make a test, and even if they do succeed in such a test, they suffer dumpshock. If a hacker is literally 'jacked in', will a second player unplugging the hacker make the willpower test an auto success, or will some further damage result? Secondly, if attacking an AI with Black ICe, it's nearly guaranteed to succeed, as the AI would lack biofeedback filters, however, neither of the 3 black programs seem to be able to affect the AI's damage tracks. Is this intentional, or are we doing it wrong?

Tracing. Against a reasonable stealth program, it seems to take forever. Is this intentional? Will a trace program narrow down the Access ID/location of the program in ever decreasing circles? What's the main point of tracing if a hacker can quickly and easily change his Access ID number?

Terminate Connection, possibly a subset of the last question, do you need the Access ID of the intruder to terminate their connection?

Deleting files on a commlink that you own is a no-test situation. Removing a file in a hostile system is not as easy (at least permanently) since the system has back-ups. Even in a hostile system it would be a no roll needed.

IIRC Black out and Hammer IC and programs don't affect programs (they attack the meat). Since an AI is not a program, well then it stands to reason these kinds of programs won't affect them.

Oh god I don't recall the detail on a trace program. THe few games I had involve Technos, and tracing them leads to a 404 error.

Ditto on the last item-but I don't think they can. They could hack you commlink and shut it doen, but they could also nuke you from orbit.

For the first question, the situation was the team's infiltrator had gotten knocked out, and the team's AI had gone into the infiltrator's commlink and was attempting to delete all of it's information before the infiltrator got captured. I didn't require a test for that, but the rules seem to imply a timeframe required to do so anyway.

Should be rare, but a search on the Resonance Realms could bring the info back. Also, save early, save often, backup your data.

Pulling the cable is auto-disconnect and normal dumpshock. Not really important why the connection was severed.


Assuming Trace rating = Targets Stealth rating, and a target unaware of the trace happening (UW pg. 104), you need to roll 30 dice on average to get 10 hits. Use hot sim: 30/(Computer skill +2) complex actions at 3 IP/turn. Should that be too slow, you can hack the target and look for location data.

Changing your access ID cuts all your subscriptions. Not convenient. Less so if you need to hack your way back in.
And yes, unless it´s "big red lever time" and innocent bystanders are of no concern. It is usually easy to get via matrix perception.

Thanks alot Ryu. Any idea on the timeframe or actions required to delete the info on a commlink?

As for tracing, I was having a spider bring up a trace ICe. Is it usually the other way around, with the spider tracing and the ICe attacking? Could I also trouble you to elaborate on



Lastly, and I realize this may be a wonky question, trace allows you to "trace an icon back to it's originating node." Would this find an AI's home node? Or merely the commlink the AI initiated it's attacks from?

Using Edit is a complex action.

Depends on the dicepools and why you want to trace. Since ICE will eventually succeed, the matrix specialist should generally do the more critical stuff.


If you are already engaged on the matrix, you don´t really care if you raise any alerts on the hackers side. Hack the attackers node - you won´t need a physical location for that. Look for map software. Ask "Where am I?". Your ICE defends the attacked node, and your matrix specialist creates a second front.


Not sure what you mean. Trace finds the originating node of the AI, not the last link of a multi-node connection.

Thanks for taking the time to explain a few things.

For the map hack, I thought you needed to actually find the hackers node first? As I understood it, the Icon/persona that appears has nothing that can be hacked, only attacked, and that a successful trace of the icon was needed to find the node. Once found, *that* can be hacked. Is this correct? Or is the trace irrelevant, and you can directly "hack on the fly" against an intruding icon?

As for the AI question, I suppose it's just a complication of the existence of a "home node". The AI tends to exist mainly in his own commlink, that a member of the running team carries on their person. For all intents and purposes, the AI only exists on that commlink. However, should he be attacked and damaged past his track, he will reassemble at the home node. I was wondering if trace would go back to the home node, or only the commlink.

I'm not aware that 'the commlink' and 'the home node' can be anything other than one and the same.

You Analyse the hackers icon. The Precious pg. 228, sidebar, first item of the list. Access ID. You hack the node with that ID. Trace gives you a physical location, which is not needed for hacking.

If you have admin rights, you can delete a file without a test. However, it is likely not completely gone. Most systems would have a backup process, and the file(s) could be restored from backup. Check the Corrupt program in Unwired for some information regarding this (pg 111-112).



Black * programs don't affect IC/AI's nor other programs.



A hacker changing their Access ID loses all of their subscriptions, so is kicked out of any systems they're in.

How the trace works (i.e. narrows down the location as it approaches the extended test threshold) is up to the GM.

Try loading the IC with it's own Stealth program; therefore it likely gets a jump start on tracing the offending icon (the hacker) before they even detect it (i.e. matrix perception).



The node/hacker/spyder just needs to have identified the offending icon via a matrix perception test, and have the appropriate priviledges (i.e. admin).

And don't forget, even if the icon logs off the system, if you have their Access ID, you can still trace them to their current location.

There doesn't seem to be anything that suggests that at all.



After looking into it further, you're (and the other kind folks who suggested it) right. Somehow we arrived at the conclusion that the connection would be jammed open, but the damage wouldn't take effect.

Thanks for the rest of the answers guys, it helped a lot!

*EDIT* Nevermind, found it.

Where does it suggest that the they can be different? Like I said, I'm not aware, but I'm happy to learn. AFAIK, the AI lives in the home node, and does everything remotely from there; that's his body, his bionode, his 'commlink'. If, as you say, "The AI tends to exist mainly in his own commlink, that a member of the running team carries on their person." and "For all intents and purposes, the AI only exists on that commlink"… that's his home node. That's the complete definition of home node (again, AFAIK).

In fact, there are several things that suggest just the opposite. For example, there are Lifestyle qualities that involve an AI taking up residence in your apartment's Home Network Node.

An AI put together less as a Hacker and more as a Rigger (Pilot origin, for example) might prefer a nicely-kitted-out drone as it's home node, for example.

I'm just trying to figure this out. The AI must have a home node. It can then load itself onto other nodes, basically traveling around (while leaving its home node 'empty'). But surely it can't "exist mainly" on the non-home node, or "For all intents and purposes, […] only exist on that" non-home node. It's just temporarily loaded there.

Pax, of course any node can be your home node, and you can change home nodes with the relevant Extended tests. What are you saying about 'suggests the opposite', and of what?

I think what the OP is saying as far as the home node issue is:

The AI has a Home Node, perhaps a CHN somewhere, but is currently “riding along” in someone’s comlink. So if traced would he be traced to his CHN (where he is not) or to the comlink (where he is)?

I’d say if he is getting the various Home Node bonuses he is in his Home Node and will be traced to there, if he is only using the stats from the comlink that he is currently sitting in than that is where he will be traced to.

I see the Home Node as literally the AI’s Home, it can leave home to work but will return there like anyone else when the job is done and it’s time to rest.

Just my 2¥
-D

Your ocmment:


.... led me to believe that at least one person thought that AI PCs had to reside on a commlink.

Oh, I dig. Thanks. I had gotten a different read, DMiller: that the the commlink (in the example) was somehow his 'primary residence' but not his 'home node'. I agree: the AI 'lives' at the home node, but can leave home to do things. And yes, the confusion didn't affect the other question: the AI is traced to 'where he is'/'where he's calling from', not to his (empty) home node.

Information regarding autonomous program's Access ID's is on pg. 110 of Unwired. Basically, they can only spoof their Access ID when logging onto a node. Otherwise it serves as a trail through the internet.

So a Track program would trace them back to the last time they spoofed their Access ID, when logging into a node. I'm not entirely sure of the mechanics of doing that, though.

Unfortunately, that section is dealing on autonomous programs, and not PC AIs. The AI section reads



It seems to make it specifically harder to change access IDs for AIs, perhaps because it's harder to trace them? (As they cannot be hit with BlackIC for the -2 threshold to trace)

Still, the problem remains as to what you would be entering if you queued up the AI's AID. As the entire persona/AI can be in the node it's intruding, it almost seems counter-intuitive to track it only to arrive back at your own node.

Well, looking at the part on autonomous programs, I would think tracing the Access ID of an AI would lead back to the node from which the AI last was in. Whether or not one could continue to trace further would be determined by whether or not the tracking entity had access to that node.

I.e. AI Bob hacks System X via a matrix connection from Runner Joe's commlink. If traced, Runner Joe's commlink would show as the origin of AI Bob's attack.

But it doesn't seem to be etched in stone anywhere. That's just my interpretation.

I seem to recall some fluff about using extraterritoriality on the matrix to 'stop' a trace, i.e. if you want to hack NeoNET node Q12E551Z3, you could first hack Renraku node B99aZ48, and through that hack Ares node T3001, all the while actually starting from your home node in a Horizon-owned public library. Assuming that you have time to slow hack some bottom-level accounts on the connecting nodes, it should provide a decent safety net against enemy spiders.

The AI has a home node, but can leave it. RW pg. 88, Welcome to My Mind.

Not having a home node is bad - RW pg. 90, Node Sweet Node. (Depending on stats, building a new home fast can be a top priority.)

Destruction of a node the AI moved to is bad. Destruction of a Home node the AI is not residing in requires finding a new home, but does not lead to destruction. RW pg. 90, Realignment and Restoration.


A trace will find the node the AI is on - in your example the team members commlink. If it works from its home node (toaster in the riggers kitchen or whatever), that location is traced. Stay "on the team" and you don´t have a long-range connection to maintain, stay home and you face less danger of physical destruction. Think "taking home in the primary battle drone" vs. "well hidden stationary home node".

.... or work from an "office" installed in a low-signature Renraku Stormcloud that can park a thousand meters over the area the meat team will be operating in (bring your Sunday Best signal rating of 6, natch).

Here's how I'm going to rule it, and I think hitting it with a fresh mind this morning helped, as I'm seeing it more logically.

The AI has to realign every night, or sleep, for lack of a better term, in it's home node. To leave the home node, it needs to access the Matrix, which it does with an Access ID. This Access ID therefore is going to link back to the home node. Now while normally a matrix specialist could spoof up an Access ID and move on from there, the process to spoof an AID for an AI takes days, making this precautionary step nearly impossible. Breaking it down logically.

To leave a home node, it has to have an Access ID.
To change an AID requires a day or more's worth of time.
If a day or more's time goes by, the AI needs to go back to it's home-node to realign.

Make sense?

The Access ID shouldn't specially link back the home node, though. It's intrinsic to the AI itself.

The AID will always link to a node, which by definition, rules out it linking to the AI.

I said it doesn't link *specially* to the home node. It links to whatever node the AI was/is on/connecting-through. The Access ID is explicitly intrinsic to the AI (same for agents, of course). So looking for the AIs Access ID would never lead to the home node, if the AI was actually somewhere else.

I'm pretty sure that doesn't really jive with the rules. When you enter an AID, you don't go to where that AID is currently browsing, you go to it's address. After browsing through the rules, I see nothing to support a node sharing multiple AIDs. Please support your statement with direct rules references.

Access ID's are either 1) the network ID of whatever device is being used to generate the persona program used to access the Matrix (for meat users of the Matrix, their commlink, CHN, or Nexus), or 2) the software ID for autonomous software (which includes AI).

Thus an AI's access ID points to themselves. The Track program simply follows their trail of what nodes they were in.

Having an AI's Access ID point to their home node means that every time that AI hacks into somewhere...they're going to need a new home node. It's very difficult not to leave your Access ID behind when you hack into someplace, as the act of logging off/disconnecting leaves an entry in the Access Log of a system that can be used for a trace. And since it's difficult for AI's to change their Access ID...

Define very difficult? I'm not meaning to be argumentitive, but every hacker I've seen goes straight for admin control, where they can wipe the Access Logs. Granted, my playing experience is limited to two playgroups and 2 con groups.

It's also not my intention to fuck our hacker over, it's just that at the moment, the AID is the only thing I can think of to use to attack him. If he runs into any kind of danger, it seems that he can log off (as Black ICe can't jam him), jump onto a commlink, set it's signal to 0 (or off entirely), and he's home free.

Gotta agree with Eratosthenes and Yerameyahu, here.

I didn't say anything about 'a node sharing multiple Access IDs'. Access IDs are your digital fingerprint, not your digital address. The AI leaves records of its Access ID on each node it travels to, routes through, or accesses. That is what Trace traces: their trail, their scent. Trace does not get the AI's Access ID and then knock on their home node. It's the same for independent agents, and it's the same for all other hackers, really: you follow the trail they left, which will lead back to the node they were using (unless they change the Access ID, which of course they will).

To change their AccesID, an AI has to make an Extended test, with an interfalof ONE DAY, not one Complex Action.


Aside from not being able to get stuck - which should be an uncommon experience for meat-body hackers too - that's no different from an Elf, Human, or whatever Hacker, too.

Meanwhile, as an AI, he has his own particular vulnerabilities and limitations.

The AID is actually more of an ICON ID rather than an IP address (which is how Almost normal is looking at it). In SR terms the device's IP address is an RTG number (IIRC-could be edition creep though).

That being said, the AI could still be tracked backed to its home node. If it moves to another node, as I recall it still has some presence in its home node.

Yeah, it's tricky in SR4 translating Access ID into RL terms like IP, MAC, etc. It's better to not try, and just go by what the book says. :/

I don't think that's the case. It 'physically leaves' the home node and moves into another node. That movement may leave access log traces (if not deleted), but there's no active 'presence' (subscription, etc.) to the home node. Theoretically, you could even move the AI to a 'working node' via a totally offline route (flash drive?).

Sorry, let me clarify. You suggested it was very difficult not leaving your AID behind. Since in my experience, most hackers wipe such traces (like most mages wipe their astral signatures), I was asking if there was something I was missing, or clarifying if you meant the +6 threshold made it more difficult.



Unwired explicitly states that any node that accesses the Matrix must have it's own unique Access ID, which then links back to the node itself. I'd provide a quote but I'm at work, gleefully talking shadowrun instead of being productive.

The home node has its own Access ID (and easier to change), because it is a node. This is totally distinct from the Access ID of the AI, which is intrinsic to it.

Great. They've wiped the acces log. But when they LOG OUT/DISCONNECT, a new entry is put in the access log with their current access ID. And since they're logged out...they can't wipe the access log. See?

It's possible, using an agent or some sort of delayed command perhaps, to do so, but an Agent will never be as skilled at editing the access log as the hacker themselves. And then the Agent's access id can't be tied back to the hacker, or it's all for naught.

A meat hacker can just spoof their access ID after logging off, killing any attempts to trace it. An AI cannot.

I'm fairly certain, but not 100%, that log offs don't register Access ID.

Could a trace action be done off of pure AID?

Unfortunately, that only adds more confusion. If an AI is riding in a commlink, and uses that commlink to hack, the Commlink would need an AID just to access the Matrix, and the AI already *has* an AID, meaning someone analyzing the AI could potentially see two AIDs.

This is giving me a headache.

Really, if you weren't confident you could handle the matrix, you shouldn't have allowed a PC AI.

I know _I_ won't, for exactly that reason: I'm not 100% on handling matrix issues during a game session. So, no PC AIs.

Well, color me purple and call me Barney.

Per Unwired, 65, it seems the Log off/Jack Out does not record Access ID.

I think I was giving too much emphasis to this (and misreading it):



But earlier it quite clearly states that it only leaves a timestamp.

That notwithstanding, all it would take is for the AI to be identified by a node/spyder/IC once, and if their Access ID leads to their Home Node...they'll be looking for a new home node.

The player knows as much (or as little) about the Matrix as I do, he's also the player least likely to be a dick about abusing rules.



A meat hacker can also be hit with Black ICe, jamming their connection on, an AI cannot.

See AccessIDs, UW pg. 48 for nodes.

Continue to UW pg. 51 for persona AIDs. Note how the unique persona AID is based on the nodes hardware AID. Just assume the AI has a two-part address of "current node AID + unique AI AID", and stuff should work nicely.

While you are at it, look at UW pg. 53-55 on Data Exchange.

Thanks Ryu, you're a game saver.

Yes. I never said they were good rules. I really don't know why independent programs have their one Access ID, honestly. What a mess.

Sounds good, Ryu.

Its not based on the real world its base on Tron. GM fiat and handwaving should be the norm, reserve tests for something important.