Subject: Re: no locale fallbacks for dictionaries (was Re: Commit: fix that spelling + im
From: Dom Lachowicz (cinamod@hotmail.com)
Date: Sat Mar 24 2001 - 21:24:32 CST
At 03:38 PM 3/24/01 -0000, Tomas Frydrych wrote:
> >I disagree. If a sentence is written in British English, then under no
> >circumstances should we try to check it with a US dictionary. If
> >you get a British document, then 'colour' is right and 'color' is
> >mispelled; it has nothing to do with Dom's (or anyone else's)
> >preferences for US spelling over UK one. If one needs to rewrite a
> >document someone else wrote in the UK to follow US conventions
> >(although why they would want to do that escapes me, since it is
> >not their original piece of work!), they have to change the language
> >from en-GB to en-US first, then they will get to see all the
> >squiggles they need, so that they can make any changes they
> >want to. If you do not follow this procedure, you will either get (1)
> >pieces of text marked as written in en-GB but spelled with US
> >convetions, or, (2) a document made of random mixture of en-GB
> >and en-US bits. Both of these alternatives, goes without saying (I
> >hope), are unacceptable.
You're quite wrong here, Tomas. Documents are intended for readers, not
authors (they're what authors produce for their readers). You're correct for
the specific case where author==reader. I couldn't care what locale someone
wrote a doc in. All I care about is what locale I'm reading it in. If I
turned in a paper with "colour" and "maximise" all over the place, my
English teacher would mark them as misspelled, even though maybe I'm from
England and they might even be printed on A4 paper.
Why one would want to is quite simple: I'm working with some people from
India and they author stuff with en-GB spellings and such. They send their
docs over, I edit them, append, and send them back. It's called
"collaboration." Also, I honestly don't care what words the author has as
custom-ignored and custom-added. If they're not in my dictionary, then
they're wrong and squiggled. Period.
Another case: I open up a german doc in a English WP. Are the german words
misspelled? Yes they are, even if lang="de-DE" *unless* I have a de-DE
dictionary. I couldn't give a crap about what langauge/locale the doc was
authored in, only the one I'm displaying it in. These are entriely separate
cases and should be recognized as such. unknown word == misspelled word in
*every* case. Unknown word in another locale != ignored word. Period.
If I have an en-GB dictionary installed, *only then* should that be used to
correctly mark the "colour" marked with the "en-GB" attribute as correct,
should that property be *explicitly* marked. If I don't, I use whatever
spellchecker is loaded by default. For me, that's en-US. For others, that
might be it-IT or whatever. Your words are then treated as misspelled, not
ignored/correct. The alternative (mark every word in an unknown locale as
correct or ignore) is horribly wrong.
This is an entirely moot point since we don't do anything with the language
tag now anyway. Well, my one patch did but no one even replied to that
email...
Paul wrote:
>NOTE -- This also implies that if we get an document of unmixed en-GB
>content, and only have an en-US dictionary available (the current case),
>then we shouldn't spell-check that document *at all*, rather than misapply
>the best available dictionary.
This argument is incorrect. You always apply the current dictionary to the
document if you don't have the dict called for by the "lang" attribute
installed.
Dom
_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com
This archive was generated by hypermail 2b25 : Sat Mar 24 2001 - 21:24:39 CST