Documentation for the Grammar Matrix Customization General Information Page
This document provides background information on how to fill out the General Information page of the customization system.
You must provide the name of the language (or grammar if you are working on a fake-language) and can (optionally) provide the language's ISO code.
The file containing language-specific implementations based on the questionnaire will be named after the name of the language. The name of the directory of your grammar will be its ISO code, if provided and else the language's name.
String processing allows you to define specific settings for tokenization, in particular regarding punctuation. The options provided on the General Information page allow you to select which characters may be part of a token in addition to letters.
The default setting discards all punctuation, except for the following three characters: "-", "=" and ":". These characters are typical ways to mark morphosyntactic boundaries and likely to occur in grammars until morphophonemics is dealt with (the Grammar Matrix supports morphosyntactic, but not morphophonemic phenomena).
The following other options are provided:
- Discard all lower ascii punctuation except \x22 (i.e. also discard "-","=" and ":")
- Split on tabs and white space, maintain all other characters
- Define a customized list of characters (other than letters) that may occur in tokens
Note that if you decide to go with the second option or allow basic punctuation marks to occur in tokens, you must either implement analyses that deal with punctuation or make sure you do not use these punctuation marks for actual punctuation in your examples. For instance: if "." can be part of tokens, tokenization will make punctuation! a word rather than punctuation in the sentence Be careful with punctuation!.
In general, the default is recommended. The other options can be used as an example on how to define the repp file as soon as the grammar gets bigger and it is time to deal with punctuation properly.
Archiving and Version Control
All information concerning archiving is optional. If not filled out, your answers to the questionnaire will not be retained by us. If possible, do allow us to save your answers and provide us with name and contact details. This information can help us monitor how the Grammar Matrix is used and improve it. We may contact you out of interest in your research, unless you add to the comments on your implementation that you prefer not to be contacted. As mentioned on the page, we will not share your answers or identifying information outside the project, unless required by law.
Highly recommended, even for smaller grammar development projects. Please consult the wiki on version control for grammar development.