On Monday 31 October 2005 18:50, Santiago Bosio wrote: > El lun, 31-10-2005 a las 20:58 +0100, Laurent Godard escribió: > > Hi Richard > > > > > Or perhaps, use the es_ES base and then only the specific > > > words that are only used in one country. Mainly > > > colloguialisms, I suppose, into a separate unique dictionary. > > > > I agree > > if the dictionary are mainly r-the same ans things can be > > unified this can save a lot of space on the servers > > > > let me know before i do the job (tomorrow) > > > > Laurent > > Ok. > > Let's clarify this because I'm going crazy here. > > Of course, they are "mostly" the same. But, for instance, affixes > are not the same for all localizations, first of all because of > variations on verb conjugations. > > I used a 'grammatical' approach, generating the affix files. And > word lists are classified by country and part of speech. > > All this work (on which I spent almost three years of free-time) > had two objectives: 1) to have a LGPL'ed Spanish dictionary, that > could ship with localised OO.org installers, and 2) simplify the > burden of generating a (future) dictionary for grammar checking.
I know and I appreciate it. I've been using a variety of your dictionary for probably a year? Has it been that long? I know I don't push the limits or perhaps even know the limits. But I ran into a similar problem after creating the various versions of the same dictionary. My port of ispell was not nearly as complete as your new version. > Now, if you remember (or search the archives), I have posted > these questions on what was the best approach: an all-in-one > dictionary, localised dictionaries, a common dictionary and > localised-differences dictionary, etc., but there was no consense > on this matter. You're asking hard questions about memory. My soft RAM is not what it used to be. Really softer. I remember the discussion but felt llike better minds would come to a better solution. I only mention this because every time I put more dictionaries on the server I get mail from Stefan asking how much more I intend to put up. :-)) So I know it is a sticky issue. If Laurent can adapt for it, if there is es_AR and es_ES then they can both be installed. Perhaps the best would be if es_AR is chosen then both es_ES and es_AR are installed and written into the dictionary.lst file, thereby getting a base language and the special, regional meanings. Would this accomplish the same benefit without too much re-working? Thanks, Richard. > So, I ask again: ¿what is the best approach? > > >From a user point of view, IMHO the best approach is this, of a > > single > > dictionary and affix file per localization. Or in the worst case, > the all-in-one dictionary as it is done now, that es_ES is > installed for all Spanish localizations. > > Unless DicOO could handle to download and install multiple files > with the same localization code (does it?), a common (base) > dictionary and a localization-differences dictionary approach > can't be taken. Because penalizes the user, who will be obligated > to do a manual install, after downloading two different packages. > > Laurent: I politely ask that you don't make any changes by now, > and that we resolve first this issue. > > Best regards, > > Santiago Bosio. > > > ----------------------------------------------------------------- >---- To unsubscribe, e-mail: > [EMAIL PROTECTED] For additional > commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
