This glossary contains descriptions of the following terms:
Annotation is detailed information about a knowledge object using structured text, diagrams and pictures.
It is often used to summarise all the relevant knowledge concerning the object and makes reference to all associated knowledge objects.
It can make use of annotation templates to structure, clarify and aid descriptions.
An annotation template is a special annotation page associated with an object which acts as the template for all descendant of that object.
Whatever is entered on the template will appear when any annotation page for a descendent is created.
Special commands in the template (enclosed in square brackets) allow automatic inserts from the knowledgebase on creation or opening of the descendant's page.
Assigning means associating an attribute or value with an object. For example, if "elephant" is an object then the attribute "colour" can be assigned to it, and also the value "grey" can be assigned to it.
Assigning can be performed in the Ladder Tool, the Matrix tool or by right-clicking on any object in any tool (except the Annotation tool) and selecting Object Attributes.
Asserting means associating one object with another object using a specific relation. For example, the object "elephant" and the object "mouse" can be related by asserting the relation "larger than" between them to create the relationship "elephant - larger than - mouse".
An attribute is a quality or feature of an object or relationship.
For example the concept CAR may have the following attributes: MAX SPEED, COST, NUMBER OF WHEELS and FUEL TYPE.
All attributes have associated values that describe the specific properties of the object or relationship.
Each attribute may be either ordinal or categorical:
An attribute matrix displays attributes and values along one axis and objects along the other axis. This is a quick way of viewing which attribute/values have been associated (asserted) to an object and a quick way of making new assertions.
Knowledge acquisition technique in which a collection of concepts (or other knowledge objects) are written on separate cards and sorted into piles by an expert in order to elicit properties (attributes and values).
Knowledge acquisition technique in which the expert provides a running commentary of their own or another’s task performance. A valuable method for acquiring process knowledge. Includes various types such as self-reporting, imaginary self-reporting, self-retrospective, shadowing and retrospective shadowing.
Communities of Practice
Groups of people performing similar roles (often in different groups and different parts of a business) that interact (often via a website) based on common needs and concerns to share and increase knowledge.
Concepts are the major type of knowledge object used in many Knowledge Engineering projects. Concepts are the things (physical objects, information, people, etc.) that constitute a domain. Each concept is described by its relationships to other concepts in the domain (e.g. in a hierarchy), and by its attributes and values.
Type of network diagram that shows concepts as nodes and the relationships between them as labelled arrows.
Type of tree (aka ladder) that shows concepts and instances and the classes and sub-classes to which they belong. All relationships must be “is a”. Also called a taxonomy.
A diagram template defines the following:
Every diagram created in the PCPACK Diagram tool must be based on a template.
The templates available to be used in a knowledgebase are created and edited using the Diagram Template Tool.
A domain is a field or area requiring expertise, e.g. physics, design or manufacturing. It is often used to refer to the area of knowledge that is the focus of a particular knowledge acquisition project. A knowledgebase will usually represent the knowledge in a domain or sub-domain.
An expert system is a type of knowledge based system that performs intelligently by emulating the knowledge and reasoning processes of a human expert. Hence, it requires a large amount of knowledge to be acquired from experts (using various knowledge acquisition techniques).
Inheritance is an automatic operation whereby any attributes and/or values assigned to an object are automatically assigned to the descendants of the object via the "is a" relationship.
For example, if the object VEHICLE is assigned the attributes MAX SPEED and COST then these attributes are automatically assigned to all the types of vehicle (e.g. CAR, SHIP, SPORTS CAR, STEAM SHIP) in the knowledgebase.
An instance is an instantiated class. For example, MY CAR may be an instance of the concept CAR.
Instances only have the attributes of their class (including inherited attributes). They may override any or all of the default values. For example, the MY CAR attribute MAX SPEED may be 90mph, overriding the default of 100mph for all cars.
In object-oriented terminology, instances are objects of a given class or type. In Knowledge Acquisition terminology, instances are real (observed or inferred) objects within a problem-solving role.
Instances are created using Ladder Down in the Ladder Tool and are related to a parent object via the "is a" relationship. They are identical to class objects with the following exceptions:
In the Ladder Tool, instances always inherit background colour information from the parent, though their font style and colour can be defined using Instance / Object Style on the object popup menu.
Instances can be shown or hidden by selecting Show / Hide Instances from the Ladder Tool View menu. A parent object with hidden instances is indicated with a greater than symbol ">" to the right of the object name though if the parent is collapsed, the collapsed symbol has precedence. An instance will be visible when created regardless of the Show/Hide setting.
Instances can be converted to Class or vice-versa by selecting Convert to from the Ladder Tool's Object menu.
The "is a" relation denotes what class an object is a member of. For example, "CAR - is a - VEHICLE" and "CHICKEN - is a - BIRD". It can be thought of as being a shorthand for "is a type of".
When all the relationships in a ladder are "is a", the ladder is a taxonomy.
Although most objects will have only one parent via the "is a" relationship, some will be members of more than one class. for example, "CHICKEN - is a - BIRD" and "CHICKEN - is a - MEAT".
NB In all the ladders created in the Ladder Tool, the "is a" relationship is the only one that operates from the child to the parent (i.e. is directed from right to left).
Knowledge is often thought of as a richer, more structured and more contextual form of information. It is required to perform complex tasks such as problem-solving, and encompasses such things as experience and expertise.
A knowledgebase is a set of files that stores the knowledge that is displayed and edited in PCPACK. The main components of a knowledgebase are:
Each of the PCPACK tools reads and writes independently to the knowledgebase. Any changes made in any tool immediately updates the knowledgebase. Hence, altering the knowledge in one tool will update the knowledge viewed in another tool.
Different 'versions' of a knowledgebase can be created, such that a new version starts as a copy of the previous version but can be altered without affecting other versions.
A knowledgebase can be published using a 'publication format' to produce a website. The resulting knowledge model is viewable on any PC that has a recent web browser (such as Internet Explorer 5.5). Hence, the published knowledgebase can be sent (e.g. via email) to experts for validation without them requiring PCPACK to view the knowledge.
Knowledge Based System
Type of Artificial Intelligence system that operates by reasoning using a large amount of knowledge, often in the form of rules.
Knowledge engineering is a field within artificial intelligence that:
For more information, click here.
Knowledge Management is a strategy, framework or system designed to help organisations create, capture, analyse, apply, and reuse knowledge to achieve competitive advantage. For more information, click here.
A knowledge object (or "object" for short) is a general term for any type of knowledge. Examples includes concepts, attributes, values, processes and relationships. Each item in a knowledgebase (as represented by nodes and links on a ladder or diagram) is a knowledge object.
A ladder is a hierarchical (tree-like) network diagram. A ladder can comprise a single type of relationship throughout or have multiple relationships.
Important types of ladder include a taxonomy (aka concept ladder) which uses the "is a" relationship and a composition ladder that uses the "part of" relationship.
Laddering is a knowledge acquisition technique that involves the construction, modification and validation of ladders. It is a valuable method for acquiring knowledge of concepts. For more information, click here.
Lessons Learnt Review
Knowledge Management technique whereby a structured group discussion takes place just after the end of a project to provide recommendations for future projects.
A link is a line or arrow connecting two nodes in a ladder or diagram. Each link represents a relationship between the knowledge objects represented by the nodes.
A marker pen (or "marker" for short) is a simulated highlighter (marker) pen used in the Protocol tool to mark-up pieces of text. Each marker pen represents a class of objects. When a piece of text is marked-up with a marker pen, it is created as a knowledge object in the knowledgebase, and is associated with the marker pen class using the "is a" relationship.
For example, if the text "vintage car" is marked-up with the CAR marker pen, then VINTAGE CAR is created as a knowledge object, and the relationship "VINTAGE CAR - is a - CAR" is created.
Marking-Up is the process of creating knowledge objects by highlighting text using marker pens in the Protocol tool. When a piece of text has been highlighted with a marker pen it is referred to as a markup.
A matrix is a type of knowledge model that comprises a 2-dimensional grid with filled-in grid cells. Important types include an Attribute Matrix and a Relationship Matrix.
The MOKA methodology provides a structured approach for developing knowledge-based engineering systems and associated documentation. For more information, click here.
Node is a term used in a ladder or diagram to refer to an element that is not a link (i.e. is a rectangular or other shaped element).
Each node represents a knowledge object in the knowledgebase.
By connecting two nodes using a link, a relationship between the two nodes is created in the knowledgebase.
Object is a shortened version of knowledge object.
An ontology is a formal representation of the knowledge in a domain. The term is used by knowledge engineers in two ways:
Within PCPACK the term "ontology" is used in both of these ways:
The term "Matrix Ontology" is used in the Matrix tool to mean the knowledge objects that comprise the matrix (axes and relation between the axes).
Ontology template (or Ontology for short) is a term used in PCPACK for a special knowledgebase on which all new knowledgebases are based. The ontology template specifies such things as:
The "part of" relation is a well-used relation in knowledge engineering to show the certain knowledge objects (parts) compose a "larger" knowledge object. For example, "ENGINE - part of - CAR", "PISTON - part of - ENGINE".
In modelling the composition of an object, a "decomposition" ladder (tree, hierarchy) is often used. This uses the "part of" relation throughout, which acts in a direction from the child objects to the parents.
Ladders in PCPACK use relations directed from the parent to the child (except for "is a"). Hence, "part of" would not be used to produce a decomposition ladder. Instead, the inverse of "part of" would be used, i.e. "has part".
Peer Assist is a knowledge management technique whereby a person (or team) experienced in a particular type of project helps a person (or team) unfamiliar with the same, or a similar type, of project.
A protocol is a term used in knowledge engineering to mean any record of behaviour, typically an interview transcript, or the transcript of the expert commentating as they perform a task. Within PCPACK, protocol is used to mean any text that requires analysis within the Protocol tool.
A relation is a type of knowledge object that represents the way other knowledge objects are related to one another. Important examples include "is a" to show classification, "part of" to show composition and those used when modelling a process such as "followed by", "performs" and "produces".
When a relation is used to relate two objects, it is referred to as a relationship. Hence, "part of" is a relation, but "ENGINE - part of - CAR" is a relationship.
A relationship is an instantiated relation, i.e. a relation that connects two specific knowledge objects.
A relationship is usually represented as a link on a ladder or diagram, or as a cell in a relationship matrix.
Matrix cells may contain a bar-symbol indicating that a relationship exists between the relevant column object and row object.
For example, for a relationship matrix based on the relation "performs", with various ROLES as columns, and various ACTIVITIES as rows, a bar in the cell that intersects the column "CHEF" and the row "PREPARES MEAL" indicates the relationship "Chef - performs - Prepare Meal".
Knowledge acquisition technique used to elicit properties (attributes and values) for a set of knowledge objects, rate them on a scale, and use cluster analysis to arrange and group similar properties and objects. Useful for acquiring detailed object knowledge and tacit knowledge. For more information, click here.
The root (also called the 'root object') is the object at the highest level of a ladder. Hence, for a ladder with single root, all the other objects in the ladder are descendants of the root.
First phase in a knowledge acquisition project in which knowledge is gathered from key experts and end users. The aim is to define the areas to be focused upon, the final deliverables and the KA approach to be used.
The SPEDE methodology is a combination of principles, techniques and tools taken from Knowledge Engineering and adapted for use in Knowledge Management. It provides an effective means to capture, validate and communicate vital knowledge to provide business benefit. For more information, click here.
Scaleable vector graphics (SVG) is a special format for diagrams used on web (html or xml) pages. When a PCPACK knowledgebase is published, the ladders (used for navigation) are in SVG format so each node can be a hyperlink to the relevant web page.
Information and skills that are not easily communicated and documented (e.g. expertise, gut feel). An expert may not even be aware they use certain tacit knowledge when performing particular tasks. As compared to explicit knowledge.
A taxonomy is an organised classification of objects according to their presumed natural relationships.
In the case of ladders, the taxonomy will most commonly be a hierarchical classification of objects related by "is a" relationships. An "is a" relationship is a relationship between two objects that implies some form of inheritance. The child inherits all the attributes of its parent unless explicitly stated.
A template is any skeletal (i.e. partly-filled) structure that applies to a group of objects and that will be filled-in later for each specific object.
The term "template" is used in three ways within PCPACK:
A value is the specific property of an object or relationship such as its actual weight or age. Values are associated with a particular attribute and can be numerical (e.g. 120Kg, 6 years old) or adjectival (e.g. heavy, young).
XML is a knowledge format most commonly used to store information in a knowledge-rich way so it can be presented in different ways (using stylesheets and transformations) on web pages.
A very common format for web pages is html, which provides a set of specific tags that identify how the text will be presented (e.g. its size, format, hyperlinks). XML can be thought of as an extension of html that allows the user to create their own tags. For example, "car" can be tagged with "vehicle" to show its class.
PCPACK knowledgebases are in XML format and can be viewed by opening them using a web-browser, such as Internet Explorer. To view the XML of knowledgebase, they are found in the Knowledgebase folder on the same drive that PCPACK has been installed.