Intersections and Search Logic

A common and simple search would be to enter a few search items on the query line, in hopes of locating a sentence which matches the idea of what you have entered. You want a sentence containing some correlation to the ideas you have entered. For example, enter on the query line:

     property evaluation

This should find a relevant response containing a connection to both entered keywords. The user would be happy to locate the sentence:

The broker came to assess the current market value of our house.

Here "assess" is linked to "evaluation", and "house" is linked to "property", within the bounds of one sentence. Therefore it is a hit, and can be brought up for viewing.

Requiring an occurrence of 2 search items is referred to as 1 intersection. The occurrence of 3 search items is 2 intersections. The default search logic is to look for the maximum number of intersections possible of all entered search items. If you were searching large quantities of text, you could add several qualifiers, as follows:

     property tax market assessment

The maximum number of intersections would be looked for; i.e., an occurrence of "property", "tax", "market", and "assessment", anywhere inside the text unit. This can be thought of as "and" logic.

To override the default "and" logic, you would enter the desired number of intersections with `@#': as "@0" ("at zero intersections"), "@1" ("at one intersection"), "@2" ("at two intersections"), and so on.

To get "or" logic you would enter "@0" on the query line with your search items. In this example:

     @0 property tax

"@0" (read as: "at zero intersections") means that zero intersections are required; therefore you are specifying that you want an occurrence of either "property" or "tax" anywhere inside of the delimited text, or sentence.

To obtain different permutations of logic you can specify a number of intersections greater than zero but less than an intersection of all specified items. For example:

     @1 property tax value

requires one intersection of any two of the three specified items. Therefore, the text unit will be retrieved if it contains an occurrence of "property" and "tax", "property" and "value", or "tax" and "value".

Any search item not marked with a `+' or a minus `-' is assumed to be equally weighted. Each unmarked item could be preceded with `=' but it is not required on the query line as it is understood. Intersection quantities "@#" apply to these equally weighted sets not otherwise marked with `+' or `-'.

Mandatory inclusion `+' and mandatory exclusion `-' logic can be assigned, and can be used with combinatorial logic. For example, you might enter this query:

     +tax -federal @1 market assessment value property assets w/page

This query means that within any page ("w/page"), you must include an occurrence of "tax" (designated with the plus sign `+'), but must exclude the hit if it contains any reference to "federal" (designated with the minus sign `-'). The "@1" means that you also want 1 intersection of any 2 of the other 5 equally weighted (unmarked with `+' or `-') query items: "market", "assessment", "value", "property", and "assets".


Copyright © Thunderstone Software     Last updated: Dec 10 2018
Copyright © 2019 Thunderstone Software LLC. All rights reserved.