Software agents are, by definition a program, but a program must measure up to several marks to qualify as an agent.

Agent Properties

Property	Meaning
Reactive (sensing and acting)	responds in a timely fashion to changes in the environment
Autonomous	exercises control over its own actions
Goal-oriented (pro-active purposeful)	does not simply act in response to the environment
Temporally continuous	is a continuously running process
Communicative	communicates with other agents
Learning/adaptive	changes its behavior based on its previous experience
Mobile	able to transport itself from one machine to another
Flexible	actions are not entirely pre-scripted
Personalization	user controls the content, function and appearance
Personality/Natural Language	agent communicates with the user in a "natural" language (e.g. chatterbot)

Agents may be usefully classified according to the subset of these properties that they enjoy. Every agent, by our definition, satisfies the first four properties. Adding other properties produces potentially useful classes of agents, for example, mobile, learning agents. Thus a hierarchical classification based on set inclusion occurs naturally.

Our Intelligent Agent technology meets all the pre-qualifications that are defined to distinguish an Intelligent Agent from a regular Software program. Our Intelligent Agent "knows" how to monitor the Web for specified information formats in regular programmable intervals, and deliver the data back in a customized manner to the requesting desktop/network. Our agent executes repetitive (programmable) tasks systematically, freeing users to focus on productive activities. In other words, the agent can be programmed to go in regular intervals to multiple information databases i.e., content sites, and inform the user of new relevant information. The agent is smart enough to know what type of information it is looking for and the results are tailored to individual specifications.

Agents also learn about the requesting user, utilizing data mining capabilities to provide a cutting edge information delivery structure, as well as user behavior prediction models that are applied to E-commerce, marketing and advertising campaigns. Data is viewed seconds after its posting and is delivered with consideration to bandwidth, availability of CPU and desktop real estate space.

Agents must be adept at interoperating with any and all technology standards, including those yet to be developed. Currently the WorldFlash platform supports Dynamic HTML (via JSP, Vignette StoryServer or one- off solutions) and XML. XML is used by some content providers to normalize the output side of an agent's request for data. Agents can generally be "trained" to parse structured information originating from any of the Internet's common transportation protocols (HTTP, FTP, POP3, NNTP, etc). Since XML is a text format it falls into one or more of these categories.

In addition to being fluent in all major Internet languages, a successful client-side agent must be designed with the target OS in mind. In this case the Windows versions were chosen for development first because of the maturity of the available tools and developer support, not to mention the excellent penetration the platform has obtained with the Internet- using population. Since the release of Windows 95, the Win32 API has been the standard method of creating efficient, smooth-running client programs. Many RAD systems have been built upon Win32 but most of them simply add tremendous complexity without delivering on their promises. For example, the WorldFlash agent needs to act as a toolbar that docks at the top or bottom of the user's screen. In order for this to happen, the Windows Shell API must be manipulated. This is somewhat tricky in any language, but the basic "C" interface is always going to be easiest.

WorldFlash is Win32 middleware that operates at the same level as the users' Web browser software. Using clever coding and combining system timers and multithreading, the agent minimizes the use of system resources needed (CPU and memory). This is critical for an agent because of the requirement for temporal continuity.