[Zaf-commit] SF.net SVN: zaf:[9712] trunk/dict/TODO

[<dwa...@us...>]

Revision: 9712
          http://zaf.svn.sourceforge.net/zaf/?rev=9712&view=rev
Author:   dwaynebailey
Date:     2009-03-23 09:54:39 +0000 (Mon, 23 Mar 2009)

Log Message:
-----------
Update TODO list

Modified Paths:
--------------
    trunk/dict/TODO

Modified: trunk/dict/TODO
===================================================================
--- trunk/dict/TODO	2009-03-20 06:49:02 UTC (rev 9711)
+++ trunk/dict/TODO	2009-03-23 09:54:39 UTC (rev 9712)
@@ -1,19 +1,26 @@
-* Pacakges can contain hyph, spell and thes with various locales
+Current TODOs
+=============
 
-OLDER
 
-* Allow different encodings for input word lists (maybe specified in the wordlist)
-* Packaging:
-	- auto create debs
-	- auto create rpms
-* Specify different ways to use words in the wordlist.  Maybe regex based.  eg.
-  Abandon words that have spaces or split words with spaces.  Or if you see
-  this char then split the word otherwise drop.  This would allow us to have
-  complete wordlists and drop what we don't need.  If the spellcheckers catch
-  up then we still have the data.
-* Look at auto creation of docs.  Eg install instructions, version numbers
-  within docs.
-* Auto upload
-* Download in cases like en-ZA that use en-GB as source
-* Allow different wordlist preprocessing rules for each spellchecker type (eg
-  MySpell doesn't care about letters or strange characters while Aspell does)
+Infrastrucure
+-------------
+Covers all checkers:
+* Drop conversion to ISO-8859-1 in build system as Hunspell can do UTF-8 natively
+* Implement negative word lists i.e. words to remove.  We have collected some words in en-ZA but haven't 
+  enabled anything in the infrastructure that will remove the words from the lists.
+* Deployments
+  * Ensure people know where upstream is for all dictionaries, check the READMEs, look at
+    OOo and other linux distro reporting.
+  * Audit and consolidate where to publish the dictionaries and how to link to them
+  * Find and burn all old dictionaries especially en-ZA with borken REP
+* Automate building of OOo dictionary extensions: we currently do .xpi and older OOo .zip we should do .oxt.
+
+
+en-ZA specific
+--------------
+* Upstream the missing en-GB words that we have discovered
+* Move from MySpell to Hunspell.  Thunderbird is our last target that uses MySpell, once 3.0 is out we can safely 
+  drop MySpell support.  There are a number of new things in Hunspell that would be nice to use, mostly we'll
+  just borrow them from en-US or en-GB.  
+* Forever: Add missing words both en-ZA and en-GB
+* Remove Americanisms


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.