HTMLArea Spell Checker

The HTMLArea Spell Checker subsystem consists of the following files:

Process overview

When an end-user clicks the "spell-check" button in the HTMLArea editor, a new window is opened with the URL of "spell-check-ui.html". This window initializes itself with the text found in the editor (uses window.opener.SpellChecker.editor global variable) and it submits the text to the server-side script "spell-check-logic.cgi". The target of the FORM is an inline frame which is used both to display the text and correcting.

Further, spell-check-logic.cgi calls Aspell for each portion of plain text found in the given HTML. It rebuilds an HTML file that contains clear marks of which words are incorrect, along with suggestions for each of them. This file is then loaded in the inline frame. Upon loading, a JavaScript function from "spell-check-ui.js" is called. This function will retrieve all mispelled words from the HTML of the iframe and will setup the user interface so that it allows correction.

The server-side script (spell-check-logic.cgi)

Unicode safety — the program is Unicode safe. HTML entities are expanded into their corresponding Unicode characters. These characters will be matched as part of the word passed to Aspell. All texts passed to Aspell are in Unicode (when appropriate). However, Aspell seems to not support Unicode yet (thread concerning Aspell and Unicode). This mean that words containing Unicode characters that are not in 0..255 are likely to be reported as "mispelled" by Aspell.

I digged the Net for a couple of hours today and I can't seem to find any open-source spell checker that has Unicode support. For this reason we keep using Aspell, because it also seems to have the best suggestions engine. Unicode support will eventually be implemented in Aspell. Email Kevin Atkinson (Aspell author and maintainer) about this ;-)

The Perl Unicode manual (man perluniintro) states:

Starting from Perl 5.6.0, Perl has had the capacity to handle Unicode natively. Perl 5.8.0, however, is the first recommended release for serious Unicode work. The maintenance release 5.6.1 fixed many of the problems of the initial Unicode implementation, but for example regular expressions still do not work with Unicode in 5.6.1.

In other words, do not assume that this script is Unicode-safe on Perl interpreters older than 5.8.0.

The following Perl modules are required:

Of these, only Text::Aspell might need to be installed manually. The others are likely to be available by default in most Perl distributions.

Mihai Bazon
Last modified on Sun Aug 10 12:28:24 2003