 Hacker Theory, Why this is possible. |
  |
 Jan 28 2007, 02:39 AM
Post
#42
|
|
|
Runner  Group: Members Posts: 2,526 Joined: 9-April 06 From: McGuire AFB, NJ Member No.: 8,445  |
Frank...
I can see and understand bits and peices of that entire thing but I cant wrap my head around it all, the last section especially. Agents, when ran outside of the Hacker's personna, run off of pre-set instructions. IE. I could tell Agent 1 to wait 20 minutes then change File 1 (the Welcome sign) out with File 2 (Elf Pr0n) which was loaded with the agent. Then abbandon the node to come back to the commlink. The hacker could hack in, drop off the agent, then go on his merry way and the agent will perform the desired tasks, based off if it's stats pass the required tests. |
 |
 
|
|
Jan 28 2007, 02:48 AM
Post
#43
|
|||||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
I think I understood everything quoted above, and agree with it, and I don't get any of the other stuff starting with the bar analogy. But if I understand you properly, you're pretty much saying that it's the TASK of running the trace system that's important to know at the "player level" and the specific method of how that task is achieved doesn't matter. We could say the Trace is being run at such a high rating because my hacker has an awesome program, or we could say its because he has thousands of little virtual-drones searching for you, or we could say it's because he's subscribed to some kind of pre-paid wireless P.I. services, or we could say he has a wireless connection to God that is telling him where you are. The point is that there are any number of ways to justify how a Trace is run because the act of "running a Trace action" is just an abstract way of describing other stuff going on in the setting. Your character wouldn't be running a Trace action, but would have other terms and processes to describe what he is doing. I can't tell if this is actually what you're describing, or just bias from my own interpretation of how agents should work because I've been making this case for Firewalls on other threads lately. Are we on the same page or reading two completely different texts? |
||||
|
|
|
||||
|
Jan 28 2007, 05:11 PM
Post
#44
|
|||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
Exactly. And it is that task that is limited in what you can juggle simultaneously. -Frank |
||
|
|
|
||
|
Jan 28 2007, 05:50 PM
Post
#45
|
|
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
But agian, why can you handle (direct, benefit, whatever) only 6 connections to 6 trace actions but can handle (direct,benefit) 1000 Database connections at a time (which is a necessity for any working computer world)?
What hinders you from giving 1000 agents the same order? Certainly this can not be restricted in any way. The point of agents is that you give them very simple orders and they do very complex task, giving you back very simple results. Its not like your node has to do anything to keep the agent going if it is running on another node. You seem to avoid that point. |
|
|
|
|
Jan 28 2007, 06:13 PM
Post
#46
|
|||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
Because the rules are mostly concerned with what a character can do, not how he does it. It's like rolling Unarmed Combat. Do we really need to know the exact vector and force behind your uppercut, or whether or not there was an uppcut at all? Before or after the dice are rolled, depending on play style, the GM or player will describe the event how they wish. What hinders a player from saying "I kick him in the gut 10 times" vs "I kick him in the gut"? As long as its one roll to hit, and one roll to damage, it works within the context of the rules, right? If players and GMs want to be more lenient in what they allow for description, the rules don't have a problem with that. |
||
|
|
|
||
|
Jan 28 2007, 06:22 PM
Post
#47
|
|
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
Only that kind of abstraction layer hasn't anything to do with SR4 matrix rules.
It's not a houserule, is't pretty much a complete rewrite... in which case, everything goes. |
|
|
|
|
Jan 28 2007, 06:25 PM
Post
#48
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Granularity. You have to decide on which level ob abstraction you are on. Once you have a system where you model single kicks, you can not prevent somebody from doing 10 kicks in a row. Once you model agents, you can not prevent somebody from ordering arround 100 angents. The only thing you can do it model the effectiveness of 100 agents vs 1 agent. And that is the path I am using in my house rules. |
||
|
|
|
||
|
Jan 28 2007, 07:11 PM
Post
#49
|
|||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
Of course you can. The Agent is an abstraction. And what i is, is a set of tasks that are being conducted under the direction of your commlink. Whether it's a network of 1000 minor processes running on 1000 nodes or a single monolithic computer program running on your own commlink does not matter. The game does not have rules for processor speeds or memory, we literally have no idea how many parallels are needed to run each individual process. If you introduce such rules (for example, the "one agent can run per device in yur PAN" that you seem to take as given), then yes, the game can immediately and easily be broken just by stacking more devices on your back. But there's no reason to add that rule. The game makes no guarantees or assumptions about where any particular program "is", and as long as you continue to handwave that point, things work fine.
Because the six trace actions are each running through 1000 database connections. Or something. Honestly, who cares? -Frank |
||||
|
|
|
||||
|
Jan 28 2007, 07:14 PM
Post
#50
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
That is excatly what an Agent is not doing. It is not under the direction of your comlink if it is running on another node. It is, by design desicion, independent. It even runs if you shut down your commlink, go away, and come back later to gather the results of teh agents task. Thats the whole point of agents. |
||
|
|
|
||
|
Jan 28 2007, 07:16 PM
Post
#51
|
|||
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
You obviously must own a different version of SR4, then. |
||
|
|
|
||
|
Jan 28 2007, 08:35 PM
Post
#52
|
|||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
That intent may exist, but it's not explicitly stated, and implementing that would be completely fucking broken. So I don't particularly care. Basically Serbitar, you are coming from the standpoint of insisting that the Agent Smith Problem is a reality and unstoppable tactic, and then confronting alternate interpretations on the grounds that they don't allow the Agent Smith Problem to overrun the world. I can't understand why you would want to do that. -- Yes, I freely grant that if you allow people to run an arbitrary number of Agents then the game is completely broken. Traces overrun all stealth instantly, unlimited IC comes to kill player characters, and Administrator Access is granted to any Hacker in one round. Right, we acknowledge this. But it's predicated on the idea that you can run an unlimited number of Agents just by having an unlimited number of commlinks duct taped together. And that's not a given under the RAW of the SR4 rulebook. It's broken, it's dumb, it'd unnecessary, and I have no idea why you cling to this notion like grim death. -Frank |
||||
|
|
|
||||
|
Jan 28 2007, 08:51 PM
Post
#53
|
|||
|
Runner Group: Members Posts: 2,526 Joined: 9-April 06 From: McGuire AFB, NJ Member No.: 8,445 |
This tells us that an agent CAN run independantly with pre-set tasks even if the owning personna is offline. What is stopping a hacker from performing the agent smith scene? The fact that the target node will notice and shut down to protect itself. |
||
|
|
|
||
|
Jan 28 2007, 09:02 PM
Post
#54
|
|||
|
Moving Target Group: Members Posts: 745 Joined: 2-January 07 From: Los Angeles, CA Member No.: 10,510 |
So do your hackers go on every run with 1000 drones fighting by the hacker's side? Or even instead of the hacker? It's the same concept. |
||
|
|
|
||
|
Jan 28 2007, 09:07 PM
Post
#55
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Because agents that work like this are a reality. Agents have pseudo intelligence by themselves, they are designed to work independently. In 2070 we will have them (in realityeven sooner). And any matrix rules worth their salt will have to find a way to cope with them. At least thats what I am trying to do. |
||
|
|
|
||
|
Jan 28 2007, 09:08 PM
Post
#56
|
|||||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
If the only defense against a Denial of Service Attack is shutting down, then Hackers have already won.
How they work is fluff, not game mechanics. Honestly, what's going on in "reality" isn't even important from a game mechanical perspective because there's a VR overlay.
I favor "abstracting the number of specific programs and the amount of processing power being used" myself. Sure, an "exploit" program probably involves having thousands or millions of separate tasks each figuring a way into your system - that's why it works at all. But allowing someone two attempts because they have "two thousands or millions of separate tasks" running is obviously crazy talk. That way lies madness. -Frank |
||||||
|
|
|
||||||
|
Jan 28 2007, 09:10 PM
Post
#57
|
|
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
Agents make bad hackers.
As they are limited to the ratings of the node they run on, they are just as good as the ICs running on it. Which basically means, instead of rolling dice, you can toss a coin. |
|
|
|
|
Jan 28 2007, 09:13 PM
Post
#58
|
|
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Well, as per RAW you just hack in with admin rights (which is quite easy even with the best nodes) and then nobody can stop you anyways. . .
|
|
|
|
|
Jan 28 2007, 09:20 PM
Post
#59
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Ok, lets rephrase it: Matrix rules in a Shadowrun universe are expected to contain independently acting semi intelligent software entities (with regard to whatever abstraction layer the rules are working on) that are essentially hackers in a box, because they have been there in older editions and because they will be there in reality. The power of these entities can be seen by looking at todays viruses, worms and trojans, which are in principle, very very dumb agents. Thats why matrix rules worth their salt will have to deal with them and to balance them. There are ways: One is to say that certain actions do not get easier, or even get harder, when lots of agents/people are doing them at thesame time (like I do in my rules with perception and hacking). Another is to say that some actions will be performed in exactly the same way by each agent so that the result for 1 agent is exactly the same as for 10 agents as there is no "randomness" (an example would be tracing). I think that this is the way to go. But thats just my opinion (and I obviously like it because it is my idea). @Rotbart: That works for Agent vs IC combat. But what about tracing, data search and other things that Frank is talking about? |
||
|
|
|
||
|
Jan 28 2007, 09:37 PM
Post
#60
|
|||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
That's all very circular though. You look at things one way and then demand that the rules also look at them the same way to be worth anything because that's how things are being looked at - but one could just as easily look at things another way and handle them differently as befits the different perspective.
If you treated every single copy of an Agent exactly the same way as a copy of a Persona in a different node, that particular proble would vanish. You'd simply treat the Agent as having a System equal to its rating and set it free along the tides of the Matrix, free to exist in ultiple nodes yet only actually acting in one per combat phase. And that would be fine. Of course, you'd still have characters with effectively limitless Medic Actions and passive scanning, tracing, and encryption good enough that the effectiveness of characters would far exceed what most people would regard as reasonable (as well as virtually excluding anything but Black IC from seriously impacting PCs. Limits on organizational effectiveness are far superior, since a limit on Agent Identity simply translates the Agent Smith Problem into a factor of wallet size. That is, a brand new Agent is really quite affordable, so the limit of one action per Agent identity is still functionally meaningless in the large bank account scenario. -Frank |
||||
|
|
|
||||
|
Jan 28 2007, 09:49 PM
Post
#61
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Yes it is. I say: A) "I want agents" (the way I defined them above). And I say B) "I want a working matrix system, that makes both hacking and defending against hacking possible". If you disregard either A) or B) my reasoning is not appropriate any more. But matrix RAW has even more severe problems: - even a single agents being better/equal to a maxed hacker - ruling everything with admin access - nonexistant baseline - extremely underpowered TMs - AR-VR problems - physically driving / rigging problems (see active drone thread) - device ratings - encryption - and more At the moment we are not having problems with agents, we are having problems with basic rules. At the moment, matrix RAW is only playable by turning your brain off and hope that your GM will fix it somehow, or playin an extremely casual rail road style way. Very bad in a player vs environment setting like Shadowrun. |
||
|
|
|
||
|
Jan 28 2007, 10:15 PM
Post
#62
|
|||
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
IIRC, both happen node-by-node, too... in which case, Agents are randomly useless. |
||
|
|
|
||
|
Jan 28 2007, 11:34 PM
Post
#63
|
|
|
Moving Target Group: Members Posts: 101 Joined: 30-December 06 Member No.: 10,493 |
Frank, that was a wonderful use of ubiquitous computing theory to justify a hacking system based on game balance rather than pure realism.
Now I can even [almost] understand why encryption is so damn' awful in SR4, because it's all using wirelessly broadcasted public keys to access data at arbitrary external adresses. And, of course, hidden mode just means you're using Tor. |
|
|
|
|
Jan 28 2007, 11:49 PM
Post
#64
|
|||
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
..please, don't mind me - I'm going to scream a while. |
||
|
|
|
||
|
Jan 29 2007, 02:39 AM
Post
#65
|
|||||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
Yeah. But a lot of those problems are fundamentally cost/benefit problems that could be solved by tweaking numbers up or down. The Agent Smith Problem is a conceptual one in which your very idea of what the Matrix is like is causing it to break. Technomancers are weak and unplayably inferior because the costs for their powers are so much higher than what those abilities are apparently worth according to equivalent character types that there's nothing nice to say about them. But the concept is salvageable. And quite easily. You just decrease the costs or increase the powers and move on. But your concept of the Agent - where it provides the benefits of an acting Hacker independently of the limitations of your character or system - that's not salvageable. The very idea is broken a priori. It doesn't matter what its stats are, or how much it costs, the fact is that it uses a quantity (money) that actual actors in the setting (corporations) have essentially limitless quantities of - and uses that to get a cumulative benefit that is used in opposed conflicts.
I cannot accept your demands because that cannot be made balanced. Your second demand is negated by your first, because your first demand rewards people for keeping track of arbitrarily large numbers of independent variables. Let's consider the core difficulty in your definition of an Agent:
That's insoluable. Why? Because no possible limit can exist for how many boxes one character could use or own. Since each one is independent, and each one is self-contained within its box, the only limitation is purchasing the boxes. Essentially, you've simply taken for granted that Agents should be equivalent to Mercenary Soldiers for hacking. And further, that unlike Mercenaries (who must be fed, supplied, transported, and smuggled), that these Agents should be available in limitless quantities stored in a virtual medium that is essentially undectable and eminently transportable. Sorry. No can do. Your idea is incompatible with a system that is balanced or even playable. An unlimited number of Agents neccessarily requires an unlimited number of die rolls to resolve. That can't function within the context of a role playing game. You can't resolve the actions of such a scenario. So you need to accept a different level of abstraction or you need to accept a computer system that you don't find "realistic" - because your ideal level of absatraction and ideal level of realism apparently involves people rolling dice to simulate each and every subunit of a multi-thousand unit parallel processor every time they do anything. That's just not reasonable. -Frank |
||||||
|
|
|
||||||
|
Jan 29 2007, 06:20 AM
Post
#66
|
|||
|
Hoppelhäschen 5000 Group: Members Posts: 5,807 Joined: 3-January 04 Member No.: 5,951 |
Thanks for letting us know that the Matrix in every Edition was broken. It's amazing how you completely fail to get the RAW, though, babbling about how System limits Programs by being the limit of simultaneous connections being handled... which are to completely different things: While the first is about how running more Programs than System degrades your Response, the second is about having not more than Systemx2 active connections. That your world of distributed processing power is not even remotely compatible with the way SR4 describes hardware ratings and upgrading them - no problem. Agents not running on your System are either in active or passive Subscription: In the first case, they count towards your Subscription limit... which is Systemx2. In the latter, they are limited to the orders they have... which is not exactly gamebreaking either, considering that they are neither sapient nor having a Persona. Wich means that they can't be at more than one Node at a time and do more than things on that Node... while being limited to the ratings of that Node. Oh, and for the record of keeping people from whining how weak TMs are - there is nothing more broken than a rating 8 Sprite. |
||
|
|
|
||
|
|
Lo-Fi Version | Time is now: 30th July 2026 - 08:29 AM |
Topps, Inc has sole ownership of the names, logo, artwork, marks, photographs, sounds, audio, video and/or any proprietary material used in connection with the game Shadowrun. Topps, Inc has granted permission to the Dumpshock Forums to use such names, logos, artwork, marks and/or any proprietary materials for promotional and informational purposes on its website but does not endorse, and is not affiliated with the Dumpshock Forums in any official capacity whatsoever.