Dictionary development and Domain dictionaries

1. Dictionary Development and Domain Dictionaries

The general purpose Gellish Taxonomic Dictionary will never be complete. When it does not contain concepts and definitions that are required in a particular application domain it should be extended. Or when an organization wants to use its own terminology, the dictionary needs to be extended as well. And when individual things are introduced as part of user data, then those individual concepts are also added to the dictionary. The Gellish language is therefore extensible with additional vocabulary and meaning, not only centrally, but also on-the-fly whenever users have a need. The vocabulary can be extended by means of the definition of additional concepts and the semantic expression capabilities can be extended by means of the definition of additional kinds of relations. Such additional definitions may be intermingled with other Gellish expressions or they may be collected in a separate amendment. The definition of extensions are themselves also expressed in Gellish.

High quality taxonomic dictionaries for various domains should satisfy a number of quality criteria.

The prime quality rule is that each concept (kind) shall be defined as being a subtype of at least one other concept, which is called its supertype concept. Such a subtype-supertype relation (also called a specialization relation) implies that the existing definition of the supertype concept is also applicable to the new subtype concept. Furthermore, the subtype concept shall be further defined by additional constraints that are expressed in a textual definition and possibly by additional relations that express what is by definition the case for the subtype concept. The specialization relation means that definitions, knowledge and requirements about all supertype concepts are “inherited” to all their subtype concepts along the hierarchy. For example, when the Gellish English Taxonomic Dictionary contains a specialization relation that specifies that a motorway is a subtype of road, then software can conclude that government rules for roads in general are also applicable for motorways. This enables that the taxonomy implies that it is superfluous to respecify the applicability of those rules for the subtypes of road again. Furthermore rules that only apply to motorways should be related to the concept motorway (and not to road). Also characteristics and possibilities are inherited, unless possibilities are constrained for the subtypes. The prime quality rule for individual things is that it is defined by getting an explicit classification relation with a concept (kind) that exists already in the dictionary.

Organizations, especially those that serve a business area, such as branch organizations, may wish to develop a complete Domain Taxonomic Dictionary, using the Gellish Modeling Methodology. The concepts in such a Domain Dictionary should then also be arranged as a Taxonomy (a subtype – supertype hierarchy) in such a way that the top concepts in the Domain Dictionary are subtypes of concepts in the Gellish Taxonomic Dictionary. This is necessary for enabling semantic verification of the correctness of Gellish expressions when using the concepts from the Domain Dictionary.

To facilitate increased world-wide development and use of a common language it is strongly recommended that required extensions are proposed as extensions of the general Gellish English Taxonomic Dictionary via this website.

2. How to define new concepts

2.1 Allocate a unique identifier (UID)

Any new concept that is added to a Gellish Dictionary or a proprietary or public Domain Dictionary shall be denoted by a new Gellish unique identifier (UID). The concepts shall be referenced in Gellish expressions by those UID’s. A new UID shall be a numeric character string that is selected from available slots in a proprietary range that is allocated on request by the Gellish language manager or it is selected from a range that is preceded by a user-organization specific unique prefix followed by a colon (:). The range 1-1000 is free for testing and querying purposes and the range 1000 – 300.000.000 is reserved. New numbers shall have UIDs that start with a prefix that is a hash (#), followed by the significant digits, optionally followed by the character E, followed by an positive or negative exponent. For example 12345E-2 to represent 123.45. A new date shall have as identifier dd:yyyymmddhhmmssp or a front end string thereof, whereas the parts of a second may consist of any decimal value.

2.2 Provide proper definitions

A new concept is properly defined and added to the Gellish language by completion of a single line in a Gellish Table. That line shall contain the following information (that express a main idea and auxiliary facts):

  • A main idea expressed by an explicit specialization relation (‘is a kind of’) between the defined concept and its (existing) direct supertype concept (or class), which expresses that the new concept is a subtype of the existing concept.

The consistent application of this way of definition causes that the concepts are arranged in a hierarchical network. That network does not have a pure tree structure, but has a network structure, due to the fact that a concept can be a subtype of more than one supertype concept.

  • A textual description that relates the concept with a text string that expresses in which respect the subtype is specialized and by what aspects it is distinguished from its ‘sister’ subtypes.
  • A (preferred) name, being an alphanumeric string that expresses how the concept is called in a particular language and within a particular language community (such as a discipline). Names of concepts that are formally added to the Gellish language shall satisfy the rules for names of concepts.
  • A language context that specifies in which language the name (and the definition) is expressed.
  • A language community that provides the context in which this name of the concept is specified as the preferred name and in which context this name is unique (within this context there are no homonyms).
  • Some other auxiliary facts that complete a line in a Gellish Expression table (this will be self explanatory when you view existing definitions in the Gellish Dictionary.

For example, a definition of a cooling system: is a kind of system that is intended to be suitable to cool some kinds of things. The concept ‘cooling system’ with this definition can be added to the Gellish Dictionary by specifying a Gellish Expression.

The core of such an expression with the definition of a concept is as follows:

UID of left hand object Name of left hand object UID of an idea UID of kind of relation Name of kind of relation UID of right hand object Name of right hand object Description
1 cooling system 2 1146 is a kind of 3 system intended to be suitable to cool some kind of things

Note that the unique identifiers 1, 2 and 3 are selected from a private range or should have a prefix. When formally added to the Gellish Dictionary the UIDs are allocated by the Gellish language manager. The number 1146 is an example of an identifier that formally identifies the Gellish phrase ‘is a kind of’ and that is part of the Gellish language.

2.3 Create Definition Models (optional)

Definitions of kinds of things can be extended with the explicit modeling of ideas about what is by definition the case for any object of such a kind. The collection of the ideas about a concept (kind of thing) that are by definition the case is called its definition model. For example, a pipe has by definition as aspect a diameter and a centrifugal pump has by definition as part an impeller. These facts are thus part of the definition models of those kinds of things. Those facts are expressed in Gellish as follows (ignoring the UID’s):

Name of left hand object Name of kind of relation Name of right hand object
pipe has by definition as aspect a diameter
centrifugal pump has by definition as part a impeller

2.4 Define roles of physical objects

Often a concept is not really a kind of a physical object, but a kind of role, such as a kind of usage or application or a kind of location of a physical object. For example, the concept ‘driver’ is a role for which some kinds of physical objects can be used and the concept ‘front end bearing’ is not a kind of bearing, but it is a bearing that is located at a front end (of a shaft). In such cases the concept shall be defined as a subtype of role or one of its subtypes (such as ‘usage’, ‘application’ or ‘located physical object’.
Thus such a concept shall be defined by a specialization relation with role (or one of its subtypes). If such a role is by definition fulfilled by a particular kind of physical object, then that fact shall be specified as well as part of the definition model, because the fact is by definition the case.
For example, the concept ‘front end bearing’ is defined in Gellish as follows (ignoring the unique identifiers and the contextual facts):

Name of left hand object Name of relation type Name of right hand object Description
front end bearing is a specialization of usage of a bearing when located at the front end of a shaft.
front end bearing is by definition a role of a bearing

2.5 Define intrinsic aspects

If the name of a kind of aspect contains a name of a kind of physical object, then the kind aspect is normally by definition possessed by that particular kind of physical object. For example, the concept ‘impeller diameter’ is a diameter that is by definition possessed by an impeller. This means that ‘impeller diameter’ is not really a kind of aspect, but a kind of ‘intrinsic aspect’, (which is a role of an aspect) in a relation with an impeller. If it is nevertheless wanted to add such a concept to the Gellish Dictionary, then the concept shall not only be defined by a specialization relation with ‘intrinsic aspect’ (or one of its subtypes), but it shall also be specified that it is by definition an intrinsic aspect of a true aspect and also that it is by definition an intrinsic aspect of a particular kind of physical object. For example as follows

Name of left hand object Name of relation type Name of right hand object Description
impeller diameter is a kind of intrinsic aspect being a diameter that is by definition possessed by an impeller.
impeller diameter is by definition an intrinsic diameter
impeller diameter is by definition an intrinsic aspect of a impeller

Note: By defining such a concept in this way the Gellish language enables computers to verify the consistency of two ways of specification: one without decomposition of assemblies and another with decomposition of assemblies. For example, a facility owner may require that for a centrifugal pump it shall be specified what its impeller diameter is, or which value it shall have, whereas a manufacturer may specify that a delivered centrifugal pump has an impeller and that the impeller has a particular diameter (value). The above specification enables a computer to verify whether the requirement is satisfied, because it can match an impeller diameter of a pump with a diameter of an impeller that is a part of that pump. (Note that one may be an imaginary pump (as designed) with UID-1, whereas the delivered real pump has a different UID-2 and those are related by the idea that UID-2 is a realization of UID-1).

2.6 Define qualitative aspects

Conceptual aspects (qualities and properties) such as colour/color, diameter, material of construction and information shall be distinguished from qualitative aspects, such as red, 6“, stainless steel and standard information. A conceptual aspect shall be defined in the same way as any other concept, such as a kind of physical object: by a specialization relation with its direct supertype and a textual definition. A qualitative aspect shall be defined by an <is a qualitative subtype of> relation with the conceptual aspect of which it is a qualitative subtype or with a more general qualitative aspect.
For example:

Name of left hand object Name of kind of relation Name of right hand object Description
color is a kind of quality that is equivalent with a range of frequencies of light in the visible spectrum.
red is a qualitative subtype of color while being similar to the color of blood.

 

Continue with: rules for names of concepts