Feature changed by: Michael Löffler (michl19) Feature #305688, revision 22 Title: system wide spellchecker (hunspell) openSUSE-11.2: Rejected by Stephan Kulow (coolo) reject date: 2009-08-12 11:30:45 reject reason: too late for 11.2 Priority Requester: Important + openSUSE-11.3: Evaluation + Priority + Requester: Important Requested by: Kálmán Kéménczy (kkemenczy) Description: Why Hunspell? MySpell is not developed for quite a long time either in Mozilla and in OpenOffice.org. Hunspell is a fork of MySpell, it has great improvements is several ways, and fully backward compatible with MySpell. Main features: * Extended support for language peculiarities; Unicode character encoding, compounding and complex morphology. * Improved suggestion using n-gram similarity, rule and dictionary based pronounciation data. * Morphological analysis, stemming and generation. * GPL/LGPL/MPL tri-license http://en.wikipedia.org/wiki/Hunspell http://hunspell.sourceforge.net/ Fedora project: http://fedoraproject.org/wiki/Releases/FeatureDictionary Ubuntu project: https://wiki.ubuntu.com/ConsolidateSpellingLibs I also find a reference in the wishlist: http://en.opensuse.org/Wishlist_Various Bugzilla: https://bugzilla.novell.com/show_bug.cgi?id=354554 Discussion: #2: Petr Mladek (pmladek) (2009-02-06 19:21:04) In general, I think that it makes sense to have one system speelchecker (disk space savings, easier maintainability, shared user-specific- dictionary between applications, ...). I do not know the implementation details but I heard from more sides that hunspell was the best choice. It seems that Fedora and Ubuntu are going to use hunspell for this purpose => we could resuse their solution OpenOffice.org has already used the system hunspell since openSUSE- 11.1. It used the internal hunspell long time before. #3: Petr Gajdos (pgajdos) (2009-02-10 18:02:44) I don't have any objections. But there are plenty of packages that depends on aspell-devel, aspell or ispell. Question is if this is real dependency or not. For example gtkspell had fake dependency on aspell- devel, because it is linked againist enchant wrapper in fact. Many of them are real, though -- they must be reworked then. #4: Scott Couston (zczc2311) (2009-02-10 21:36:17) Apart from the purely technical aspect, Myspell offers very very poor English US and slightly better English UK suggestion choice and is hugely dependent on the user getting the first two letters correctly spelt speelt spellt in the word that is mispellt. - English is such a poorly constructed language when it comes down to conventions that always apply. There is very little logic in the English Language so without * Improved suggestion using n-gram similarity, rule and dictionary based pronounciation data. * Morphological analysis, stemming and generation MySpell is very poor. - The other notable drag with the very very poor UK dictionary that still gets confused between US/UK words - and there is a huge difference. Quite often I cut and past into Open Office as the spell checker and word choice offerings are just so much better - in every sense. I would love to see us use the same spell checker and language modules that Open Office use so we can stop relying on other dictionary modules being developed and make use of the continual development that is done for us by OO. - Love the 1 dictionary, love the continuity idea, but rather than reinvent dictionary modules can we not use OO and its spell checker process. #5: Petr Mladek (pmladek) (2009-02-11 11:35:57) (reply to #4) I see that we have started to use hunspell in OpenOffice.org by default since OOo-2.0 and January 2006. I am surprized that there is such a big difference between myspell and hunspell but... In each case, it looks like you have a good experience with hunspell. #6: Scott Couston (zczc2311) (2009-02-11 22:10:48) My PC always has a copy of OO Writer Open - I write all the text anywhere else and before I commit I cut and paste to OO then correct spelling and cut and paste back into dialog - Its that bad - and the worst thing is the enormous amount of personal dictionaries any use can have - I think we are up to about 7 or so user specific dictionaries counting OO ;-) Other advantages are there is an abundant number of localised .LEX(?) files available for OO/FireFox/Thunderbird #7: Scott Couston (zczc2311) (2009-02-27 23:19:23) The Other Wonderful thing is that if we make this change the enormous amount of dictionaries that are available in OO and continue to be developed by Mozilla and other Interested Parties is a function win in every respect - We will never have to re-invent the wheel in both the type of spell checker and its associated dictionaries. The shear amount of localisation between English in a UK/US/AU form is so unreal and the size of the dictionary is so much larger than the current. Whilst we are about it - Could we also have the ability to *easily* remove a word that was added to a dictionary - I don't know about you - but getting out a huge text editor to find the word we accidentally add to a dictionary of any sort, is the biggest time waster of them all and we end up never getting rid of our additions that are commonly created by clicking on the wrong button during spell check. - NOT spell as you go - Spell check of completed text. While we are at this can we go to the point in providing a Universal Network Dictionary Server NDS - I think that's best for me to open a new feature request - but lets get this up and running much better in 11.2 -- openSUSE Feature: https://features.opensuse.org/305688