21 Spelling checker

The Thot editor allows the user to check the spelling of a document being processed. Checking is based on dictionaries [Richy92a]. The checker compares the words of the text with those in the available dictionaries. It suggests several replacements for each word which has not been found in the dictionaries. The user can then choose the correct word from the list of suggested words (or can type another word). The replacement operation is performed by the checker.

The spelling checker can correct multiple errors in a single word. The error probabilities are used to order the proposed corrections (see section 23.5). Thus, the proposals that correct a single error are placed before those that correct multiple errors; errors in accents (a e instead of à ê for instance) before deletions or permutations of letters.

The checker can work on multilingual documents: it consults different dictionaries according to the value of the Language attribute (see section 6.1) applied to the text being checked. Dictionaries are currently available for French, English, German, Spanish, Dutch, Swedish and Italian; dictionaries for other languages can be added (see section 23.5.1). The parts (or documents) written in languages for which there is no dictionary are not checked.

In addition to the language dictionary, the checker consults the spelling dictionary specific to the document. This dictionary has the name of the document followed by the .DIC suffix. However, the DICODOC configuration parameter (see section 23.5.1) allows you to choose a different dictionary name. By defining this variable, you can use the same spelling dictionary to check all documents. This dictionary is created and automatically updated by the spelling checker when the commands ``Skip (+dic)'' and ``Replace (+dic)'' are issued (see below).

The ``Tools'' menu, at the top of the document windows, contains the command ``Spelling checking''. This command displays a form which enables you to choose the correction parameters and to interact with the checker:

• The maximum number of proposals is specified in the ``Number of proposals'' area. This number can take any value between 1 and 10, the default value is 3. Once an error has been detected, any change made to this value immediately displays a new list of proposals.
• Several buttons allow you to ignore certain words: for example words that contain capital letters, digits, Roman numerals or special characters indicated in the list below. This list can be modified by the user.
• The ``Correct where'' submenu, similar to that in the ``Search'' form (see 3.8), allows you to specify the part of document on which the checker will work.
• The erroneous word is displayed in the area ``Spelling checking...''. The list of proposals is displayed in the selector located below the incorrect word. By default, the first proposal of correction is selected. To select another proposal, simply click on it. The selected proposal is displayed below the selector. This proposal can be directly modified in the form. The correction is made once you have clicked on one of the buttons ``Replace and next'' or ``Replace (+dic)''.
• The language in which the correction is made is specified in the top left corner of the form. To change the language, just apply an attribute Language to the text concerned by using the ``Attributes '' menu and restart the correction procedure.

To start the correction, click on the ``Skip and next'' button at the bottom of the form. If an error has been found, the incorrect word is selected and highlighted in the document. If no error has been detected, the ``Not found'' message is displayed in the bottom right corner of the form.

The user can then use one of the buttons placed at the bottom of the form to perform one of the following operations:

• ``Done'': the form disappears and checking is aborted. Previous corrections are not canceled.
• ``Skip and next'': the word considered incorrect is acceptable. No change is made in the text nor in the dictionaries. The checker searches for the next error.
• ``Skip (+dic)'': it is the same as ``Skip and next'' but the word considered incorrect is added to the spelling dictionary specific to the document, thus the checker will ignore any subsequent occurrence of this word.
• ``Replace and next'': the word considered as erroneous is replaced by the correct word (changed or modified by the user, see above) and the checker searches for the next error.
• ``Replace (+dic)'': it is the same as ``Replace and next'' except that the replacement word is added to the document dictionary.

[Section 22] [Table of contents]