User Tools

Site Tools


This is an old revision of the document!

The Gellish Expression Format

The Gellish Expression Format is a tabular format. It can be expressed in various basic formats, such as CSV, JSON or spreadsheet formats, such as XLS, typically using the UTF-8 encoding standard. The following columns in the table form the core of the expressions, using English terminology and provided with an example of a Gellish expression:

31 101 3 201 7 4
Name of an intentionName of a left hand objectName of a kind of relationName of a right hand objectSymbol of unit of measureTextual definition
assertionThe Euromastis located inRotterdam

The above Gellish expression table is language dependent and human readable. It can be made language independent by adding references to unique language independent identifiers (UIDs) by extending the table with columns for such UIDs as follows:

1 5 2 60 15 66
UID of an ideaUID of an intentionUID of a left hand objectUID of a kind of relationUID of a right hand objectUID of a unit of measure

Note that

  • The UIDs shall be distinguished from the numeric UIDs that are standard in Gellish. Users can make their own UIDs by a prefix followed by a colon (:), followed by free codes. In this example Column 1 with UID therefore contains the prefix 'pr' and code '101', resulting in UID pr:101.
  • The UID of an idea is intended for being used for making statements about the expression as a whole.
  • An 'assertion' is a standard kind of intention with the Gellish UID 970025.
  • The Euromast and Rotterdam do not appear in the Gellish Dictionary (assumed), thus the user can allocate his own UIDs for those concepts.
  • The kind of relation 'is located in' is a standard phrase for the standard kind of relation with UID 5138.
  • The unit of measure column is not applicable for this expression, thus the columns for the UID as well as the symbol is left empty. If the columns are not required in a whole table, then the columns can be deleted from the table.
  • The Textual definition column is intended for a human readable definition of a concept. It is intended to be used only on a row where the concept is introduced by a classification relation ('is classified as a') or by a specialization relation ('is a kind of'). As this is not applicable in the example, it is left empty.

Multi language support

There is a separate pair of columns available for specifying the UID and name of the language of the name of the left hand object on each row. This enables to use various languages in the expressions in one table, including the specification that term is a translation of another term for the same concept. Furthermore, one or more separate columns can be inserted each of which for specifying alternative names for the left hand object in a specific language. The column ids for those columns should be the Gellish UIDs for the particular languages. For example, a table in English may include an expression in Dutch and an additional column with a name of the left hand objects in German (where applicable). This is illustrated in the following table:

54 31 101 910038 3 201
Name of a languageName of an intentionName of a left hand objectName in GermanName of a kind of relationName of a right hand object
EnglishassertionThe Euromast is located inRotterdam
DutchassertionDe EuromastDie Euromastis a translation ofThe Euromast


Alias names for objects, such as synonyms, abbreviations, codes, etc. can be specified by explicit statements using phrases for the kinds of relations that express the appropriate subtype of the alias relation. There is no extra column required for such expressions. Aliases are specified in a similar way as the specification of a translation in the above example. For example:

Name of left hand object Name of kind of relationName of right hand object
PC is an abbreviation ofpersonal computer

Homonyms (language communities)

Using the same name for different concepts, being homonym names, is enabled by distinguishing the objects by their different UIDs and by specifying different language communities in which those names find their home. This is supported by a special pair of columns in the expression table for the UID and the name of the language community. For example the term 'bank' in a the language community 'business' denotes an object with UID 990152, whereas the term 'bank' in the language community 'civil technology' denotes an object with UID 700140. This is illustrated in the following core table:

16 101 3 201 4
Language communityName of left hand object Name of kind of relationName of right hand objectTextual definition
business bank is a kind oforganizationthat is intended to provide financial services.
civil technologybankis a kind oflandthat is located alongside the border of a water.

Contextual facts

Other columns are available for expression of contextual facts, such as status, date of creation, author, etc. Those columns can be added depending in the requirements of the user by selecting them from a list of available columns. The columns may be arranged in any order that is convenient for the user. Multiple tables can be combined and different tables may consist of different collections of columns. This is described in detail in the document 'The Gellish Syntax and contextual facts'

Header rows

The table has three header rows. The first header row contains the following fields:

GellishEnglishVersionversion codedate of releasename or category of expressionsfile name

followed by an optional sequence of name based parameter names and values:



  • The string 'Gellish' is an obligatory standard term in the first field.
  • The name 'English' in the second field may be replaced by another language name, such as 'Nederlands'. It specifies the language that is used in text fields, although on each row it is possible to overwrite this by specifying the language that is used for the name of the left hand object. Furthermore there can be one or more additional columns that specify names of the left hand object in a particular language (as is described above).
  • The following five fields are optional free text fields.

The name based parameter names and values are intended for use with automated UID generation only.

  • The optional Lower_obj_uid and next three UIDs may specify ranges for new numeric UIDs.
  • The optional prefix specifies a prefix that may be used to precede a colon (:) and newly generated numeric codes for generating new UIDs for new concepts. For example, the prefix 'pre' can be used by software for generating the following sequence of UIDs within the object uid range 20-30: pre:20, pre:21, pre:22, etc. The software should first verify which is the highest current value within the range for the prefix.

The columns in a table are identified by language independent column IDs in the second header row. The third header row contains free text names of the columns, corresponding with the column IDs.

Various examples of tables in Gellish Expression Format are given in the download area.

A detailed specification of the definition of the tabular format is given in the document 'The Gellish Syntax and contextual facts' that is available in the download section of the gellish website.

==⇒> next page

gellish_expression_format.1540920393.txt.gz · Last modified: 2018/10/30 18:26 by andries