Dnia środa, 23 marca 2005 11:00, Bram Moolenaar napisał:
> > > The current mechanism is a bit inefficient. But it's fast, and we
> > > need it to be fast. Loading the spell file can be a bit slow
> > > though.
> > If I open 4 Vims with spellchecking on everything will be slow.
> > I tried with normal ispell data - it generates 38M Vim dictionary.
> Are you sure this is a correct dictionary?
> For English the "'s" item,
> which can be added to any word, almost halves the number of words.
> Perhaps something similar can be done for Polish. It's not perfect
> though, since it allows something like they's.
No. I lack of dictionary (;) to explain it in extenso but in Polish it
can not give good results - too many endings, too many declinations. In
many places it could treat quite common mispells as proper words.
> > One more thing about loading time: loading of dictionary for Polish
> > takes 3-4s. For mails/usenet articles it is HUGE penalty - and this is
> > most often use for spellchecking.
> More than a second to load a dictionary would already be undesired.
> I suppose using the engspchk.vim script would be even slower for you.
Yes. And 3-4s is for minimum (11M) dictionary.
That's why use of some interface ispell/aspell/myspell would be good
thing. They don't require loading of dictionary.