Re: PO-based Documentation Translation



Hi Folks,

On Thu, 2003-10-02 at 09:47, Karl Eichwalder wrote:
> Malcolm Tredinnick <malcolm@commsecure.com.au> writes:
> 
> > One item I forgot, but just remembered re-reading that stuff: the KDE
> > tool requires every string to be translated before a translated document
> > can be generated

I'm not sure this is a good idea - we have similar problems in Sun where
a particlar message cannot be translated (perhaps due to an
internationalisation issue, or maybe a convention like where syslog only
contains english messages for example)

The mechanism in gettext for falling back to a default english message
if a translation is not found, is far better I believe and was designed
for this purpose.

The reason I think it's bad, is because it has the potential for
screwing up other translations down the line.

For example, imagine having a "translated" (but still in english - hence
dirty translation) message in a Japanese field in the translation
memory.

Later on, perhaps this english text gets reused as a "translation" for
the next occurrence of that same message in a different context (where
we could display the localised message) is nasty and difficult to spot,
because according to your TM, "All messages have been translated
successfully"

Of course, you could get strict about it, and require exact matches from
the previous version of that message file only, but then you're loosing
the advantage of having a huge database full of strings...

 	just my $0.02

	cheers,
			tim




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]