You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Sebb (JIRA)" <ji...@apache.org> on 2017/04/16 18:54:41 UTC

[jira] [Commented] (TEXT-78) Investigate adding a spell checker

    [ https://issues.apache.org/jira/browse/TEXT-78?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15970489#comment-15970489 ] 

Sebb commented on TEXT-78:
--------------------------

This is likely to involve a lot of work, even if only English is supported. And ongoing maintenance.

Also I think it's starting to stray from the Commons mission.

> Investigate adding a spell checker
> ----------------------------------
>
>                 Key: TEXT-78
>                 URL: https://issues.apache.org/jira/browse/TEXT-78
>             Project: Commons Text
>          Issue Type: New Feature
>            Reporter: Bruno P. Kinoshita
>            Priority: Minor
>              Labels: help-wanted, spellcheck, spellchecker, spelling, text
>
> There are already spell checkers such as [Hunspell|http://hunspell.github.io/] (which has its own API to define suffixes and prefix), and simpler spell checkers like [jazzy (forked)|https://github.com/reckart/jazzy].
> Hunspell (GPL/mixed) is written in C++, and Java programs that use it normally use native access through [BridJ|https://github.com/thomas-joiner/HunspellBridJ]/[JNA|https://github.com/dren-dk/HunspellJNA]. Jazzy (LGPL) is good, but not actively maintained, and simply looks up words from a dictionary.
> We can provide a simple interface, and perhaps one or two different spell checkers in pure Java.
> Having edit distances already in the project, we could investigate something such as adding a [BK-Tree|https://en.wikipedia.org/wiki/BK-tree], an in-memory look-up spell checker, etc.
> When creating the initial interfaces & design, it will be important to remember that we want to support multiple languages too.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)