One of the differences between a dictionary and other kinds of (scientific) texts resides in the fact that a dictionary is structured both to a high degree and in a uniform way. Dictionary making is less of a creative than of a diligent and, sometimes, stubborn and tiring activity. The classical motto for composition, variatio delectat, is absolutely inappropriate in lexicography.

Inconsistency implies that part of the information given is wrong or misleading and that the user is hindered, rather than helped, in orienting himself in the vocabulary.

The following are aspects of consistency among the components:

• There is a well-defined metalanguage which is set out in the appropriate main sections and stuck to religiously in the entry lists. For instance, there is a biunique mapping between the conjugation classes defined in the grammar section and the conjugation classes used in the grammatical classification of the entries. If wing is semantically classified as ‘animal body part’, then claw cannot be classified as ‘body part’.
• Definitions of related entries complement each other to exhaust their lexical field. They bring out relations of hyponoymy, antonymy etc. If chemistry is defined as a science, then physics cannot be defined as a scientific discipline, etc.
• Definitions must not be circular.

Structural analogy relates to aspects such as the following:

• All the entries are specified for the same kinds of information.
• The order of components in an entry, and their typographic marking, is always the same.
• Senses of polysemous entries are always ordered according to the same principles.

The computer is an invaluable aid in securing consistency. The more of all the procedures involved are automatized, the more one may be certain that their result will be consistent. To give a trivial example: In the printed dictionary, the grammatical category of each lemma is to be typed in italics. If one were to apply this font modification manually to the grammatical information entered in the corresponding database field, one would be liable to make mistakes. The proper method is not to do this manually and instead to add the italics automatically to the content of this field in the moment that it is exported to a print file.

As already mentioned, an important tool in securing consistency in the form of the content of a certain field is the range set.

A proved technique of securing consistency in the microstructure and the content of the fields is the following:

• Select all the database entries of a given category, e.g. all the transitive verbs or all the body parts.
• Make the DBMS display a list of entries that shows a subset of the fields (other than the one used for the selection).
• Focus on the content of some field, for instance ‘morphological structure’ or ‘meaning’.
• Check whether that field has been filled in always in like fashion.