Content-Transfer-Encoding: binary

To err is Human, to really foul things up requires a computer.

The HumanBeing is incredibly complex. The dream of artificial intelligence comparable to human or animal intelligence is still a distant dream. We finally have isolated MachineAgents, robots, that can walk with near animal agility, and others that beat us at chess, but most of the thousands of skills that comprise living animals still elude automation.

Thus HumanAgents are necessary in in much of our activity and when they aren't necessary, they are often desirable.

We comprise those human agents in a SocietyOfMind that comprises our CollectiveIntelligence.

Agent: any thing that does something
HumanAgent|: any person that does something

In the context of WikiWorld, an agent is any thing that improves, refines, organizes, archives, adds to the collective thought process. I see at least two classes of agents in operation at WikiWorld; HumanAgents and MachineAgents.

Defining a human agent requires definitions for the terms human and agent. Defining human may exceed the scope of this page.

A human is an end in itself.

An agent is a means to an end.

See HumanPrincipals, [[|http://www.halfbakery.com/idea/Pringent#1001874306 Pringent]]

==Discussion ==

====Use the term 'Actor'

==

David, it sounds like what you are looking for is 'Actor'. A non-passive something. That is the term applied in Unified Modeling Language|, and generally accepted in most software design. %%% %%% An 'agent' is merely a subset of an 'actor', as an 'agent' is a software service that acts as an actor with a limited interest for a human or human system. For instance, an Agent that tracks your favorite TV shows times, and is capable of suggesting new views based on the profile of what you've told it is your favorite type of shows. Or an agent that checks all your email accounts, and alerts you to email from your Significant Other and auction updates (you've given it a hot list of sources to alert you about when emails from your hot list arrive in one of your various email accounts). Agent is a term used exclusively in software for non-human actors. :D Humans are users, even if they are your travel agent. And until a conversational agent can pass the TuringTest, software will not confuse the two. --StarPilot

'What if we considered the entire travel agency a complete system? In this case the user of the system would be an individual seeking to book a seat on an airline. The system would then be an object with several interfaces some for input and some for output. The travel agent would simply be an advanced user interface for the user of the travel agency system. User inputs criteria for acceptable flights; travel agency system outputs list of flights with most acceptable on top. User selects and approves best choice; system creates ticket. User inputs payment; system outputs ticket. The travel agent would be a user of the software used to track bookings and such.%%% %%%The user, however is not concerned with how it calculates its output, just that it happens in an efficient and user-friendly manner. More importantly in the context of the ObjectWiki project, the developer or analyst of such a system (one that incorporates HumanAgent (s) with MachineAgent (s)) would need these terms to adequately describe these systems.%%% %%%Now the question: If all the pieces and parts of a travel agency are considered one system object, could the travel agent be considered an agent of the system? %%% %%% Thanks for the input StarPilot. I will take a look at the Unified Modeling Language| for more semantic guidance. I actually considered ''actor' but intuitively 'agent' felt better. Most processes (services?) are completed by humans using machines as tools. I would consider word processing to fall in this category. Service:Create Minutes for Staff Meeting Agent:Individual using word processor. In other words the word processor doesn't really do anything per se. Now a service within a word processor might suggest a better spelling for a word. This active spell check continually scans the input for collections of characters that do not appear in its list of approved words. This is a service that actually does something. This MachineAgent may even change the word as you type with out waiting for input from the user. I suppose I need to take a closer look at the UML for a better, more accepted structure for my thoughts. Why reinvent the wheel, right? --DavidSiegel

Viewpoints, David, viewpoints. From say, my viewpoint when I'm looking to book a trip up to Boston to visit family, the travel agent is just that, my travel agent. I make a request, she does her thing, and presto, I have my travel itinerary. From my viewpoint, the whole travel agency and travel agent are one system.

From her viewpoint, I'm a customer, but she is the user interfacing with her company's system. Customers come in, she handles the input and interactions with the agency's systems.

From the company's system's view point, she is just an actor. There are no customers, merely users.

I, as the customer, 'Act' on her by making my request. She translates my requests and 'Acts' on the system. The system processes her requests, and presents her with the various choices. She translates those choices into terms I can understand, and we negotiate the process.

We see that I am an 'Actor', as is she.

However, from the travel company's system view point, we can clip me out of the picture... I'm an external second layer 'Actor' to it who only acts on that agent, its 'actor'. Therefore, I can be simplified away in that view, as only the agent using the interface is important to it.

'Actors' initiate change, in state or merely data values. Without an 'Actor', the system is static.

Agents tend to be associated with Actors which posses some form of 'Intelligence'. Sometimes, MachineAgent would be appropriate, for instance, in the case of a computer process that scans a stock prices stream, and initiates a sell order if the price reaches a high or low value. However, when an Actor is nothing more then an oscillating input wave, no intelligence is associated with that input, therefore it would not be called an agent.

Actors initiate change, in that point of view. Agents are a subset of Actors that have some form of 'Intelligence' associated with it. Users, contrary to what you might think, do not have any form of Intelligence presumed to be associated with them. They are considered to be randomly flailing away at the input surfaces.

Consider... Gravity is an Actor on Mass. Our Sun's light is an Actor on our planet's ecosphere. In neither case, do we associate intelligence with either gravity or sunlight. My stock broker is a financial agent acting in my interest (in certain cases, anyways :-D). The difference? Intelligence.

Humm... You know, Intelligence is just the process of differentiation. So... an Actor that is meaningfully (to the process involved) differentiating can be said to be an Agent, and one that is not is only an Actor.---StarPilot