Open the Editor via the screensaver control panel.  

Open the editor by double clicking on Nav8.exe executable.  

Open an agent module into the Editor by double clicking on an agent module .XO file extension executable.

How do I modify an Agent to search for something I want? (and)

Each agent supplied with the software can be modified by opening the editor Nav8.exe. The editor opens with a graphical display panel topped by a toolbar of icons.

To start with, look over the brief documentation on the particular agent by clicking on the -?- icon in the toolbar.

Now open the main agent Search set up panel by clicking on the –X- icon. The existing agent will have various settings already in use.

Save-As the agent module to a new name (or Save-As-Empty, - only the network is thrown out, the settings are retained), and then re-open the Search set up panel for the newly saved agent module.

AND/OR line keywords: In modifying an existing agent for use at the same web site and for the same basic purpose, all you need to do is look at the word terms present in the –ANY of- and -ALL of- text lines (which are the Boolean OR and AND lines respectively) and enter your own terms.

If there are no terms present, and just one term is present, like ‘GMT’ in the -ANY of- that term may be used to cue collecting text from a certain point in the document. In that case do not modify it.

Metric channel word counts: The next step is to change the word terms present in the Metric A, B and C text entry lines to be words that fit the subject or topic you are searching. The terms do not have to be present in the web pages you search and the three Metric channels A, B and C are used to count word incidence in the individual web pages searched. Enter your own terms into each Metric.

Use each Metric channel as a category of words, such as politics, economy, and education for instance. You would place words related to politics in the first channel, words related to economy in the second, and so forth.

Once done, and after you have run the agent and collected a network of web pages and related word count (or numeric count) statistics, you can instantly regenerate and re-rank the Report results according to each different metric, or a sum of all the metrics. This way you get to focus on different categories of results within your original search which you may have initiated from Google or ABC-News, etc.

For example, you could search BBC News web site for news articles. Your search term you entered in your browser at the BBC site might be ‘terrorism’. You will get a lot of news hits on that subject unfortunately.

If you have entered a group of sub-topic words into the Metric channels, you can sort or rank the articles from the BBC according to those sub-topics. Subtopics could include a group of words typed into Metric A for domestic Homeland security, words typed into Metric B relating to armed forces activities abroad, and words typed into Metric C could relate to alerts, threats, and breaking news.

Number extraction: You can modify the agent to collect a number in a particular location on the screen. For brief documentation details on how to use this more advanced feature click on the -?- in the upper right hand corner of the agent Search set up and tracking window (click on the –X- icon on the far left of the main graphical display toolbar).

After clicking on the -?- a question mark will appear attached to the mouse cursor.

Now click on the text entry line for Metric A, and that will show you instructions on how to set up different numeric counts.

Text sentence extraction: You can assign words you have entered into the Metric A line for example to be used to only collect sentences from the web page that contain words from that line. It works on a left to right basis, so place your most important keywords you want to use to key text sentence extraction, on the far left of the list.

The Beta Surf3D release only collects up to a 1000 bytes of text per web page, and so the additional words in the Metric A list may not be used to key text sentence extracts if the first words have used up the 1000 byte allotment.