[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-i18n-doc
Subject:    Re: Strawman Proposal: Smart typo correction to reduce burden on
From:       Albert Astals Cid <aacid () kde ! org>
Date:       2009-03-01 21:52:23
Message-ID: 200903012252.24989.aacid () kde ! org
[Download RAW message or body]

A Divendres, 27 de febrer de 2009, Andrew Coles va escriure:
> > -1.
> >
> > I disagree, fixing even just a typo can change the
> > "understood meaning" a
> > translator of can get from a message (seen that happening
> > in the past), so
> > always manual check is needed.
>
> Okay, let's see if we can find useful cases where typo corrections don't
> need checking.  A quick yes/no will do.
>
> 1) Use of British English
>
> - Before   "Configures the Behaviour of X"
> - After    "Configures the Behavior of X"
>
> 2) Splitting contractions
>
> - Before   "Changes won't be saved"
> - After    "Changes will not be saved"
>
> 3) Common mis-spelling
>
> - Before   "An error has occured"
> - After    "An error has occurred"
>
> 4) Missing . at end of message, in cases where a message contains several
> sentences.
>
> - Before   "Here is a message.  Please read this"
> - After    "Here is a message.  Please read this."
>
> 5) Exclamation point usage
>
> - Before   "The file %1 cannot be found!"
> - After    "The file %1 cannot be found"
>
>
> As you can see, some of the changes I make on a day-to-day basis are quite
> pedantic - perhaps we're interpreting 'typo' differently ;).  So, my take
> on these:
>
> 1,2,3) Doesn't need checking by translators
>
> 4) Grey area.  If the translators left off a ., to match the syntax of the
> original string, then this might prompt them to add one.
>
> 5) I'm out of my depth, perhaps if ! has implied a more forceful
> translation it wouldn't be reasonable?
>
> > KDE 3.5 is really really really frozen solid, please drive
> > your typo fixing
> > efforts to KDE 4.
>
> Oh, I am doing, but I'm also thinking about how I can ease the burden on
> translators having to check that cannot means the same as can't etc.  As
> people reading this list will well know, translators are a valuable asset,
> so anything that can help... :).

I'm with Pino here, you can break more than you can fix, so i'd vote for no 
automated scripting in this case.

Albert

>
> Andrew


[prev in list] [next in list] [prev in thread] [next in thread]