I guess the second option is the most useful, for all languages, imho. Maybe it could even be expanded to even increase to edit distance 3, but limit it all to x alternatives found or x ms.
It seems a lot to increase time from 16 to 20, but it depends on the amount of cases it is needed for. Maybe test it running on a huge dirty corpus with just speller rule? Ruud > Hi, > > we currently don't offer any suggestion for 'Cincinatti' in English, as > the correct word (Cincinnati) has a distance of 2 from the misspelling. > What could be done about it? > > 1. Use a maximum spell checking distance of 2. This decreases > performance from 16ms per sentence to 24ms per sentence (with some > Wikipedia text I used for testing which had some foreign words in it). > > 2. Use a maximum spell checking distance of 2 only if the spell checking > distance of 1 doesn't find anything. Performance still decreases from > 16ms per sentence to 20ms. > > 3. Add a replacement pair for this very case, but miss all the other > cases we probably don't even know about. > > Did I miss any other solutions? I'd tend to go for solution 2. > > Regards > Daniel > > > ------------------------------------------------------------------------------ > Want excitement? > Manually upgrade your production database. > When you want reliability, choose Perforce > Perforce version control. Predictably reliable. > http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk > _______________________________________________ > Languagetool-devel mailing list > Languagetool-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/languagetool-devel > ------------------------------------------------------------------------------ Want excitement? Manually upgrade your production database. When you want reliability, choose Perforce Perforce version control. Predictably reliable. http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk _______________________________________________ Languagetool-devel mailing list Languagetool-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/languagetool-devel
