That is a decent explanation for a web server, however I cannot find anywhere in the RAW that implies that a server/client data connection still exists or is used. Everything in the RAW is done by moving your persona from one node to the next. That means a logon request, getting through the firewall, and then accessing the node. There is no "Get Node Webpage" free/simple/complex action that allows for that kind of direct access transfer of data.

Maybe I'm just missing something obvious here, but my impressions were that the data was not sent back to your Commlink, but to your persona via the persona's icon. It is the persona which is your interface to the Matrix and the icon is your physical location on the Matrix. Your commlink is just a node where you jacked in from. Heck, it may not even be a Commlink, but could be your residential terminal, or a terminal at the Corp your breaking into.

The text from the RAW that I quoted stated that a node can have public access areas or even be completely public access. That reads to me like it is saying that an area of the node is open to public access, and you need to connect to the node to access that area. If it were saying what you seem to think it says, I would expect it to say "Date/Icons/Programs/whatever on a node can be publicly accessed". You are saying that the information on a node can be publicly accessed remotely (from another node), if setup for it. I think the raw is saying that the node itself can be public accessed if setup for it, and that information on the node can be accessible from that public access area.

Its been my, and I think cetiah's, assertion that just cause your icon crashes does not indicate your commlink crashing which you seem to agree with in your most recent post but disagree with in the post quoted herein.

Okay, if this is your opinion, then the whole discussion between you and I is just a misunderstanding. You use "comlink goes offline" and I use "you are disconnected from the Matrix" but it sounds like we're saying the same thing.

It started with me trying to explain to Frank that a Persona and your comlink's Node are two very different things. A persona can be crashed through cybercombat, but nowhere does it say the node also crashes.

I mentioned that your node can still be hacked even though your persona was crashed in cybercombat, and that was the point where you objected. Apparently, you thought I meant hacked from the Matrix, which wouldn't make sense if you weren't connected to the Matrix anymore. I meant just wirelessly. Your persona is crashed, but your comlink is still functioning and its node can be hacked (locally, not through the Matrix). So the node stays unaffected while the persona is crashed. Thus, the persona and the node are two different things... which is what I was trying to explain.

But I think we both agree that if you are defeated in cybercombat, then nothing from the Matrix can hack you because you are not connected to the Matrix. But my argument was that the node is still there when the persona isn't. And independant IC operating on that node (that is, IC independant of your persona) still continue to function on that node.

Really? I'm kind of dreading that. As Frank pointed out in another thread, it will most likely be contradicting the core rules (and maybe itself) in pretty significant ways.

I'm not arguing against agents or mobile code. My example above includes an agent moving from system to system. With that stated, I am abso-fragging-lutely unwilling to accept that public hosts will alow unknown code execution from rank and file users. It would be java in reverse, with the host executing code provided by the client. It's an administration nightmare and barring some earthshaking concept of worthiness that should immediately be patented, I do not accept it as plausible.

I find idea of it in today's world to be somewhere between foolish and ludicrous. I really wouldn't accept it in SR 2050 when mobile code (virus) destroyed the net in Crash 1.0. I most definitely, undeniably, and irreconcilably refuse to accept it as plausible post Crash 2.0, when mobile code anhiliated the net again.

SR4 has sufficient computing power in the world that there's no need to allow users to run agents on their systems. Only particularly overworked people will have the need for their agents to run on something other than their own Comm. Users can have agents do all the work from their comm while they sleep/eat/get their groove on.

And yet its here. It's happening. Deal with it.

(As I said, there are security issues. But no one's saying, "Uggh. Security issues. Let's just give up." There are techniques and tricks to verifying remote code, and other suggestions on how to improve network security in a mobile-host environment including having an environment that limits the functioning of these agents. And the benefits for mobile-agent networks are extraordinary. Take a read through those links. It's fascinating stuff.)

Your persona is a remote represenation of your comm/workstation/etc. It can exist on multiple remote systems simultaneously according to RAW (p.118, I believe, in the section aptly called "accessing multiple systems.")

However you only have one persona, it is shared across all remote hosts. Given that you only have one persona while being on multiple systems and that your persona's attributes are based on your comm/workstation/abacus rather than the remote server, obviously the persona is running on your local comm while maintaining a client/server connection.

As for the "get node webpage" action, it's unnecessary since it is rolled into the concept of the VR interface. You say "connect to GlobalDynamX" and, ta daah, you are in the public portion of their matrix site. It will have some VR metaphor, probably their office or something, and you can pick up the virtual documents (read the "web pages") or go through the virtual security door (with some username/password challenge) to the private system. The matrix is essentially VRML but before VRML was a term anyone knew.

Your Persona is a particular bit of software that runs on some piece of physical hardware. The Hardware in question is usually your comm and virtually always in immediate physical proximity to you. I say "virtually" always because some joker will come up with a thousand foot long data cable or wifi "cantenna."

The Persona is essentially your avatar, an icon with a particular "skin" that represents you. The matrix is very similar to a game of Quake. You have a default environment (home map) as your theme. You connect to a remote server. Your computer actually renders the environment to you as well as other people. It bases the actions of the other people on the information relayed from the host server, which is itself basing the other people's appearance & actions on the information provided by their computer.

I agree that information on the node can be accessed by the public. I agree that it is possible for someone to create a completely public server, intentionally or otherwise, that allows remote code exploits. I've known of intentionally wide open systems that allowed anything to be done to/with them but they were CD-based machines with no logs that rebooted nightly for the explicit use of people with questionable motives.

I disagree that it is likely, plausible, rational, or even probable that it would be the case in actual implementation outside of the Shadow community.

I'm familiar with remote agents. I think they have value...eventually. But my base arguement is not specifically security but that it is unnecessary. In today's world it's being used for Grid computing and other high-end applications where more CPU power is necessary and it is worthwhile to have code that essentially self-tunes itself.

Why, in 2060, when everyone has the equivalent of a Fairlight Excaliber in their pocket, would it be necessary to have code migrate from server to server? What advantage does it give to either the agent-owner or the host owner? What is the advantage to Google? What is the advantage to Fuchi?

IF there's no advantage, why would anyone put up with the (surmountable) security hassle? Give me a justifiable advantage that isn't solved easier and cheaper by providing a company comm with a higher rating.

"Consider the case of a traveller wishing to arrange a trip from a town outside London to a city on the East Coast of the USA. This is the sort of activity – surely – that agent researchers promise people will be able to delegate to their personal travel software agents, in the same vein as we do to our human secretaries. Today, it is still largely the case that the secretary would consult other human travel agents, who in turn contact yet others to arrange your flight and itinerary. These others include hotel agents, railway agents, rental car agents, etc. However with much of the information now being found online, but being owned by different stakeholders who all want to make money from the information and service, the promise is that your personal travel agent (PTA) will negotiate with other software agents representing the interests of the different stakeholders. This way, your itinerary gets generated with minimal or possibly without any human intervention, unless changes to the itinerary are required. This is truly a multi-agent problem wherein the inter-operation of separately developed and self-interested agents provide a service beyond the capability of any agent in the set up, and in the process all or most gain financially. Economically speaking, all these agents have comparative advantages over each other due to specialisation, and trading their services is good for all. This lofty goal generates some critical challenges that the hype of software agents and many papers in the literature cast aside. We describe some of the main problems next, and comment on how far the challenges they pose have been realised."


This was taken from the following website, a paper on "The Potential Benefits of Software Agent Technology for British Telecommunication Laboraties, by Hyacinth S. Nwana and Divine T. Ndumu for their Applies Research and Technology department.

(The author advises against agent research, by the way, saying there has been very little real progress in this area.)

That's not Serbitar, it's just his agent checking in with us. In my custom house rules we call them 'rats'.

Valid reasons. Now tell me why the code has to EXECUTE on those systems as compared to on a CPU that you own.

I can write a stupid agent. It will read web pages looking for keywords, follow links, download documents, etc and run on my PC the whole time. I have friends who have Ebay snipers to purchase items for them during the frantic last minute bidding process.

Tell me what is the value in having it execute on other systems. Don't tell me why agents are good, I already know. Tell me why anyone would let me run my agent on their hardware. Tell me why I would need my agent to run on their hardware.

Alright. You win. I can't respond to that without opening up whole other areas of discussion (introducing ideas that have not only been disagreed with before, but "officially" shot down as not existing in SR).

So what was your point? I thought it was that agents are bad. Is it just that RAW is bad? Or Tron-like VR is bad?

(It's still a fascinating read, btw.)

My assumption is that Serby is following posts that he finds particularly agreeable with \signed to indicate that he agrees, and/or would have said the same thing.

Just a guess.

Oh. No.
"Access" in RAW means you have an icon there. For your persona to access a node, it must have a persona icon there. That's how firewalls protect a node; they prevent the persona from entering.

A hacker access multiple nodes has multiple icons. That's quoted in the section you are referring to.

Independant agents exist as independant icons. They, too, travel to any node they want to access.

Further, only personas can access multiple nodes. It never says anything about agents. Presumably, this is because agents have to be loaded onto a particular node to act independantly of a hacker's persona.

\signed

It's those damn rats again. They're such a nuissance.

Hehe.

\signed = I agree.

Exactly. Is all about expressing your opinions. I really hate it when worthwile posts do not get the attention they deserve.

And you guys have pretty much got me agreeing with you that it should work this way, assuming the topography of the Matrix being the way it is and computers more-or-less working the way they do now.

I have no objections if you want to modify RAW and say it works this way.

Yes and no. Roughly from my 20+ readings of the rules since starting my participation in this thread and the amazingly dedicated discussions prompted by various posters here, this is the final "basic breakdown" I've come down to:

System = CPU
Response = overall speed, various hardware and software factors included
Persona = active memory
Node = harddrive
OS = OS. Sort of. OS is run off the node. OS runs the Persona. OS loads programs from the Node into the Persona.
AR/VR = a method of presenting icons
icons = user interface, or GUI
persona icon = the "active workspace" or "active window" that the user is viewing icons through.

Nope, Personas work this way but agents don't. According to RAW, independant agents are nothing more than independant icons moving around through the system and do not have a corresponding Persona. Agents do not work the same way as personas, and must be loaded onto a particular node to act independantly (i.e., conduct actions there without your persona). They cannot access multiple nodes at once; only Personas can. They cannot project icons onto other nodes; they ARE icons.

Personally, I think this is an oversight in RAW and creates sooo many problems. It would have made more sense just to say Agents have Personas, too. I guess this was to prevent Agents from loading other Agents; I don't know.

Also, as far as I can tell, Agents can't load programs. Or rather, they can't activate programs to load. Any programs an agent is carrying/running must be available in the comlink and loaded into a persona. Agents can also be loaded into a persona so that the agent can employ them. The agent can be loaded up with a copy of the program, and thereafter if the agent becomes independant it can continue to run the program, even if there is no longer a persona actively running that program.

I think it's all very complicated, but its RAW. If you want to house rule this stuff away, I don't mind.

there's no need to modify the RAW. that's how agents work. agents simply have an option that hackers do not: they can change what node their persona is operating on, by uploading themselves (and presumably deleting themselves from the previous node, though that's not required).

the point of the "must be uploaded" rule is so that it's clear where an agent's code is running. this is a change from SR3, where it was completely unclear where exactly an agent's code was being stored/run at any point in time. it doesn't necessitate having the code running on the same node it's acting on, it simply necessitates having the code running on a node somewhere.

Agents don't have a persona. That's nowhere in the rules.

Independant agents are icons. IC are icons. IC must be loaded into a node that it is protecting. Whereas a hacker who breaks in could use all the agents loaded in his persona that travel with him.

Your persona is a representation of your comm/workstation/etc, however it is your Persona's [b]icon[b] that can exist on multiple remote systems simultaneously. The matrix section of the RAW often makes a distinction between the Persona and the Persona's icon.

Your Persona is running on your local commlink managing its icons and passing along commands. However the persona's icon is really on their node and executing programs. This gives the Persona control over the connection and the node doesn't have to really care about the details of the connection. If it was the server/client connection you want, then the node would control the connection.

The advantage of this is that since the Persona controls the connect, the persona is limited to the number of connections it can have (System*2) and the node doesn't because it isn't sending data anywhere but a local icon (in this case).

While a server/client scenario may be more likely, plausible, rational or even probably, it isn't supported by the RAW. The RAW continuously talks of the persona being our connection to the Matrix, and your persona's icons being on a node and hoping to other nodes. There is an entire section on logging into nodes and everything you can do and manipulate and account access levels.

There are no sections which detail how to access servers on a node. None on how to exploit a server to bypass the firewall so you don't have to use a logon action. Nothing on anything which suggests a client/server scenario.

Also consider the following: Why does it take Black IC jams open a connection? In the Persona/Persona Icon world, the Black IC is jamming open the hot simsense connection between the icon and persona. It has to be done through an attack sequence because the persona controls that connection, not the node. If the node were acting as a server then it would control that connection and what gets sent over it, so why couldn't any IC just immediately send Black IC level simsense down the pipe? Why doesn't IC just immediately drop the connection altogether, since the server controls the connection?

So what? They are still running on a "particular node" - which means that the difference between them having a Persona and not is purely semantic. They still have their particular node, and they spawn icons in the nodes that they are accessing, and their stats are capped out by the node they are running on, and if you crash their icon they collapse and are booted from the Matrix.

I don't understand why you have this big hangup over whether Agents technically have Personas or just have a particular node - it doesn't actually have any effect whatsoever on what happens when an Agent is accessing 6 different nodes.

-Frank

Why do you let people install spambots on your PC?

Oh, that's right, you didn't let them, but they did it anyway. That's what attack agents are. They are compromised systems that the attacker controls to some extent.

Because we believe--and we believe that the RAW supports us in this-- that an agent cannot make multiple connections, and access multiple nodes, but must instead move from node to node.

Also, remember, the Matrix of 2070 barely resembles the internet of today. The Matrix of 2070 has been through two complete, ground up rebuilds. it is not a simple evolution from what we have now to then.

Seriously, what the hell? Agents can access nodes in addition to moving from one to another. It's an explicit statement under "Using Agents".

I mean holy crap. This thread is 6+ pages long. The Agent section is less than half of 227, and less than half of 228. You could at least read the whole thing.

-Frank

I did. And that quote you provided should go "Agents can also access other nodes independently if instructed to and if they either have the passcodes or are carrying an exploit program and can hack their own way in (as independent Icons)" While I admit this passage is rather confusing (it uses "and" twice with no commas.) it seems clear that accessing other nodes includes hacking their way in as independent Icons.

Of course, if I had just read your quote, and not read the context, I may have come to the conclusion you seem to have.

Allow me the opportunity to see if I can shed light on what you consider is Pyritefoolsgold's and my confusion. To do this I will walk you through my train of thought on reading this section of the manual. We can't correct any misconceptions unless you know what angle I'm coming from. =)

An expanded section from the manual that contains the statement you quoted reads:

I'm going to go line by line of the above section.

"Agents use the Response attribute of whatever node they are run on; this means that the attributes of an agent operating independently may vary as it moves from node to node."

This statement explicitly states that agents have the ability to jump from node to node. It even states there are affects on the Agent for doing such a thing. This does not directly state that agents cannot access another node, but it doesn't mention the ability to either.

"Agents can be loaded into your persona like other programs (taking a Complex Action), allowing the agent to accompany you to any nodes you access."

Here we see that because the agent is loaded into the persona, it gets to tag along as the persona accesses another node. Lets take a second to look at the login process a bit. One of the persona's icons does a logon action to access the node. The persona then combats the node's firewall. If successful <or if the persona knows the passcode> then the Persona has access to the node and uploads its icon. It must upload its icon, since the icon gets on the node and so the persona icon moves from one node to another. The upload occurs as part of the successful logon.

"Agents can also access other nodes independently if instructed to and if they either have the passcodes or are carrying an Exploit program and can hack their own way in (as independent icons)"

Well, since we want to keep things simple, we assume that agents behave as much like Hacker as possible. Also remember that the agent is itself an icon. So the agent want to access another node. It does the logon to the node. The agent then combats the node's firewall. If successful, the next step would be for the agent to upload its icon. The agent's icon is itself, so the agent icon would be uploaded. So what happens here? Well with the hacker, the persona icon moved from one node to the next as part of the upload, so the logical leap <at least in my mind> is that the agent icon moves from one node to the next as part of its upload. This does not leave the Agent accessing two nodes, but has the agent hoping from one to another which we know is possible.

My apologies to Pyritefoolsgold if this is not what he was thinking with his statement, but these are indeed my views on the matter.

Sorry, man. It's the rules as written. You're inventing stuff so that agents act just like personas, when they don't. And it seems like you're doing it just so you can bash how RAW works when it doesn't work that way. You're the one who insisted we only stick to RAW and not discuss house rules or how we would like to change the Matrix so that it makes sense.

You are taking the quote out of context. "Access" means access that node which requires loggin in and entering that node, and in some cases hacking. "Independantly" means independent of your Persona. Agents are either loaded in Persona or are operating independantly (meaning they are loaded in a node as an independant icon).

Well, what do you expect? You've ignored our attempts to quote rules to you. The game text defines how agents work and the game text defines how personas work, and you insist on attributing qualities of the persona to the agent even though that's nowhere in the game text. You just keep repeating the same thing, not to quote RAW, but to specify your own conceptual interpretation about how RAW should read if it worked the way you seem to want it to. And then you accuse us of not reading the text when we quote passages and systems to you. We talk about how different functions work and how different elements of the game fit together and how this fits with the overall topology and gameplay and nothing works... no matter how much RAW we quote to you, you just present us with your overall "So what? You're stupid." non-argument.

I suggest you go back and do a little reading, beyond just two or three paragraphs because people here have been quoting text that talks about personas, icons, logging in, cybercombat, etc. that all relates to reasons why you are wrong on this issue. Maybe if you read more than just those few sentences, the reasoning would be more clear to you. Actually, I'm not even sure you're reading those sentences because the only thing you seem to be reading at all are the three words "agent" "access" "independantly" and then you attribute your own definitions to those words, take it all out of context, and insist that the picture you are left with is RAW.

no, it's not the rules as written, it's the rules as you've interpreted them. you have decided that agents must be loaded onto a node in order to hack that node, and you're interpreting every ambivalence in the rules to support that. FrankTrollman, myself, and others, have decided that agents make sense, and are therefore interpreting the rules in such a way that the make sense.

Because the only reason Frank and Serbitar are arguing for this position is that it creates their Agent Smith problem of another thread and they can turn around and declare how unreasonable and stupid RAW is because of what (they believe) it says.

They want agents to be omni-powerful "hackers in a box" ready to unleash themselves on any node in the system, and that's not what RAW says or why those rules were written.

Their interpretation also pre-supposes contradictions in the rule system where they don't exist.

And according to Serbitar's original post, this thread was meant to convince certain people who disagree with him that RAW does actually work this way and everyone who believes differently is wrong. That's the challenge that's been presented in this thread. So it ultimately doesn't matter which side is "right" because no one's trying to fix the problem, if one exists. The purpose of the thread is to see who is right so that corrections, interpretations, and new ideas have a framework with which to launch in other threads.

Right?

Ancient Fucking History on a Pogo-Crutch, people!


You've filled two threads about f'nuckin' Agents faster than they filled one thread about statuatory rape over on the Wizards.com (im)mature forum.


Why don't you just write the publisher and ask them for a clarification?

Hey, I wanted to propose a set of assumptions for a good Matrix system, discuss them, and move on from there, but Frank insisted I take down my post because it wasn't appropriate for this thread and he was only interested in discussions about RAW.

Although Serbitar started this thread and he should get the final vote on what is or isn't within the spirit of this thread, he's kept silent and he seemed to be more or less satisfied when I admitted that RAW needs to be changed and house rules could cover the issues much better than RAW. Since Serbiter's issues have been more-or-less covered, anyone should have the right to say "let's take the conversation this way now" and Frank did that. I consented, and no one has contested the issue.

So per his request, this topic is all about RAW and deciding correct interpretations of RAW, regardless about how we feel about the issue or whether or not we feel these custom hacking rules do the job better or not.

Since the person who started this thread has his own Matrix rules, and so do I, it seems you would be wrong, Garrowolf. We are not arguing about this because of our limitations to see past RAW. We're arguing about this because its become clear in other discussions that not all of us seem to be starting from the same viewpoint when discussing changes and ways to move past RAW.

They don't NEED personas. It is the software that translates your commands into computer commands, aka a User Interface. Software doesn't need a user interface. It speaks computer natively.

IC are agents. All software and data have icons; icons are just symbols/images/animations that provide something to see in VR. Your icon online is provided by software, referred to as a Persona.

Agents have inherent iconography programmed into them to represent themselves on the net. Their attributes are based on their own codebase (which, since it can run other programs, is an OS unto itself) and includes their own firewall rating.

The main difference between an Agent and a Hacker is that if you crash the Persona you crash the node but an agent can crash without impacting the node. That's a vital difference since IC are agents and you don't want a server to crash just because the IC died.

Your persona can generate an icon on multiple hosts and yet your Persona still resides on your CPU, not that of the host system.

It's more appropriate to say a Persona imitates an Agent than Agents imitate Personas. A persona is required to provide the interface between human thought and the matrix. Agents are natively matrix entities. There is nothing to suggest that a persona can do anything an agent can't other than act as a user interface.

An excellent theory.
So what are you saying? Agents should be equal to personas? Worse than? Better than? None of these interetations are supported by RAW. RAW just says that they are different and has different rules to cover them.

I didn't say they HAD icons. The quote says they ARE icons. My very point is that they don't HAVE icons and can't project them.

Also programs don't exist within the Matrix until they are loaded into a node, persona, or agent. So they might have an icon, but that icon is still loaded into something -- your Exploit system isn't going to jump a node by itself. But if you load it into a persona, the persona can carry it. This is how programs move from node to node within the Matrix. They have to be loaded into something. They can't just "project icons" like a persona can.

No, you're own definitions instead of RAW. Independant agents *are* icons and they can carry programs. Their attributes are based on their codebase and the node that they are loaded into - this also supports the passage that an agents attributes can change as it moves from node to node. Nice try.

Why do people keep saying this? It's not in RAW anywhere! You're node does not crash if the persona crashes! In fact, nodes can't even crash. They can be shut down, but they can't be crashed. Only programs and OSes can be crashed, and crashing the OS is one way to shutdown the node. Your node is not shutdown when your persona is crashed.

Yes, but this is a function unique to personas. Agents cannot do this. It specifically says that the agent must be loaded onto a node, it can't just be run off the comlink like a persona.

This isn't supported by RAW at all. It isn't even reasonable to assume from a game design standpoint. The agent was made with certain limitations and I believe these limitations exist for a reason - to prevent agents from being better than personas, i.e., better than hackers which is what Serbitar and company object to in the first place.

I'm with you there. That's true.
But if you want to extend the analsys to, "...therefore, agents are better." Sorry, RAW doesn't say that and the rules clearly give the edge to personas.

There's a bunch of stuff listed that personas can do. But you choose to read "personas or agents" whenever you see personas listed in the book or in this thread, so nothing I can say is going to change that.

Some rules apply to agents, and some to personas. Agents are not the same as personas. You shouldn't assume rules apply to agents just because they apply to personas.

Wow, here I was worried that I was going to take the 'most senseless expression of exasperation' award away from Frank for this thread, but you clearly topped us both. Good job.

Umm... yay?

Why would the publisher care? Do you think that would help? Maybe they play Shadowrun, too! Of course, knowing my luck, he'd just quietly chuckle to himself and go back to reading his original digital copy of Unwired.

it doesn't change the Agent Smith problem. you can still have an unlimited number of agents helping you. they just can't multitask, your way.

I know, I know. I just couldn't resist. I thought it was funny, and I found your comment more or less unhelpful and insulting so I posted an unhelpful insultful comment, too. Not the most mature thing in the world, I know, but I haven't had any coffee since 6pm.

I don't do that for two reasons:

1) Bothering writiers of a book after its been published is just... tasteless. Maybe tasteless isn't the right word, but I wouldn't want to be bothered by such things if I was writing for fanpro. Especially if I have to take time away from writing/creating stuff to answer disgruntled fanmail and clarifications. WIZARDS did a good job in hiring a full-time guy who did nothing but this for their magazine publications, but its unreasonable to ask people to do it who really aren't being paid to do so.

2) Ultimately, everytime anything "official" seems to come out for the Matrix rules it contradicts what's in the book and no one is left satisfied. The FAQ discussion on agents' programs for example pretty much directly conflict with what's in the book. Also, developers/writiers almost never want to post their reasoning behind their decisions leaving more folks disgruntled and confused than would otherwise be the case. Finally, if the writers/developers post online material or emails that can be considered "official" and this material isn't in the book, then the learning curve for new players coming into the game just skyrockets. No, it's better if all "official answers" come from a book or regular errata updates.