Quoting Dmitri Gabinski <[EMAIL PROTECTED]>: > Hi!
Hi Dmitri, Many thanks for your testing and bug report. I will soon revise and fix these bugs. > > I've tried out HunSpell on OOo 1.9 m126 / Windows XP and must say, I'm > generally disappointed about it. > > First, however, I'd like to point out that HunSpell does have the > advantage of supporting UTF-8. I've successfully checked spelling in > Moldovan (Romanian Cyrillic). Unfortunately, that's the only plus I > can see. > > Now about minuses: > > A) All languages: HunSpell produces much less suggestions for > misspelled words. And that means, that in some occasions you don't > find a variant you'd want to use. > Here's one example in Russian. Instead of �Фи�Па� (�Firm�, > �Company�) > I write �Ф�иПа�. > HunSpell suggests for that case: > ��иПа > ��иПа > ��иПа > Obviously, none of that variants is what I need. > MySpell in turn suggests: > п�иПа > С�иПа > г�иПа > �и�Па > ��иСа > ��и�а > So, there's the right variant, although MySpell ignores the fact that > the word should be written with the first letter capitalized. Myspell/Hunspell has a heuristic ngram suggestion algorithm for `very poorly' spelled words. Hunspell has an improved classification (see test data: http://qa.openoffice.org/issues/show_bug.cgi?id=35725), but it seems, also has a too small/strict default suggestion number sometimes. I will fix this problem with a more flexible suggestion number. > > B) Russian: HunSpell ignores the letter ���. One explanation: that > letter is not obligatory in the modern Russian orthography and many > people normally use �о� instead. I am among those that insist on using > ��� and I have the ru_RU_yo.* module for Russian in OOo. Again, one > simple example: I edit a text with �о�-spelling and have, say, > �оо� > instead of �о�� that I prefer. �оо� is marked as error and > MySpell > suggests �о�� among other variants. HunSpell does not. Can it be > connected with the fact that the Russian module is based on the KOI-8 > encoding? Perhaps this problem is the same as C). > > C) Belarusian: HunSpell counts words at sentence beginning (i.e., with > the first letter capitalized) as errors and suggests lowercase-only > spellings. That's not acceptable. Myspell in OOo 2.0 beta has a new cp1251 character table with fixed casing data. I will update old MySpell character tables in Hunspell, too. > > D) Esperanto: HunSpell seems to be unaware about that language. At > least, I can see no HunSpell option for Esperanto in the Writing aids > menu of OOo. This problem (using old dictionary manager code) has solved with the integration (a screenshot: http://hunspell.sourceforge.net/esperanto.png). I will also fix it in the UNO modul. (Esperanto has a special eo_ANY locale. `ANY' is supported only by the new dictionary manager.) > > Considering all of the above, I must say that although HunSpell is > theoretically very good, in practice it in its current state can't > serve as MySpell replacement for OOo. Of course, HunSpell already is a > good option for languages that MySpell can't handle. It seems, Hunspell lacks some new Myspell beta bug fixes and improvements. Fortunatelly, it's not too late to fix. I compare and update Hunspell with Myspell of OOo 2.0 beta till Wednesday. Please, could you make an issue for these bugs with test data? I will attach your issue to Hunspell CWS for Quality Assurance, too. (Only tested and approved CWSs are integrated into OpenOffice.org.) Thank you for your very useful bug report! Best regards, Laci > > Best regards, > > Dmitri Gabinski > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
