ATypI Dictionary: brief technical manual

This dictionary is like a paper dictionary, adding the usual benefits of the web such as searching. However, it is structured rather differently inside. Before editing anything it is important to understand the structure, so that your changes have the right effect throughout the system.

Languages

The dictionary is in many more languages than a paper dictionary. As well as languages, there is (optionally) country-specific data that indicates where a word is only in use in a particular country, and there is also support for regional usage. When a language is specified, you can also specify the country and region if necessary, or leave them blank if the word’s usage is common across most countries and regions in that language.

Different definitions can be provided in different languages and countries (e.g. US English and British English), but not currently in different regional accents. The differences between definitons in different countries should be minimal. For example, in the British vs. US situation, they should usually be limited to spelling changes.

So: words (headwords, in dictionary jargon) must have language, and may have country and region attributes. Definitions must have language, and may have country attributes.

Objects

Normal dictionaries are based on words. A word has one or more meanings, with definitions for each. This dictionary is based on objects. An object is not a word, it is the meaning of all the words that refer to it. For example, ‘£’ is an object. Although there is a word for it (‘pound sign’), it is important that we are talking about ‘£’ as ‘pound sign’ also has other meanings (Americans use it to describe a ‘#’). Objects can’t usually be represented literally in the system. Some (like ‘£’) can be represented by their Unicode character, because they are characters, but the scope of the dictionary is much wider than that. I have picked ‘£’ because its object can be represented in this manual. In the current dictionary, this happens to be object number 927, but the object numbers are not displayed to normal users.

Definitions

Objects have definitions. For example the definition of ‘£’ is something like ‘Currency symbol for British Pounds Sterling. It is derived from initial L in Latin librum’. They can have many definitions, one for each language (At the time of writing, there is also a Spanish definition of ‘£’ – ‘Signo de esta forma: £’). The system does not enforce the number of definitions that a word has, so it is possible for an object to have no definitions. It is important to realise that definitions should not include information such as the history of a word, because they do not describe words, they describe things. For example, a defintion can say ‘This item was invented by Guttenberg in … ’ but not ‘This word comes from the latin … ’. Note that in the ‘£’ example, it is the object itself that comes from the latin, not the word for it. The current version of the dictionary does not allow the etymology of each word to be stored.

Words

Finally, objects have one or more words that can be used to describe them. For example, ‘£’ has the word ‘Pound sign’. It is important to realise that the word comes last, not first. This is because more than one object can have the same word. For example, ‘Pound sign’ is also an American way to describe a ‘#’, but ‘#’ is a different object (it is object number 179). All the words for the same object are synonymous, although they may be in different languages. It is possible that certain objects will not have a word to describe them in certain languages. In that case, only the definition will be available.

Words can be marked as in current usage, or obsolete. A word being obsolete does not imply that the object it describes is obsolete, only that the object is no longer commonly described with that word. It may be that the object is completely obsolete, but the word to describe it is still current. Currently, this is the only etymological information that is stored, but this part of the system may be improved in later versions if there is a demand.

Robert Munro