Hi,
pardon me for jumping in.
It's plain personal interest on the subject in general (l10n/i18n) and some expertise and practice as a copy editor (for German), and a creator of English->German translations.
My 3 cents:
First of all, literal translations barely make sense, that's why I don't (really) understand the fuzz about why some word's "literal equivalent" doesn't make sense if it's translated, well ...
literally -- of course not!
Translations ought to be done
by meaning and context. Exclusively. Period.
Culture and habits of the target language/country must be honored as well. Any braindead attempts to do a word-by-word translation are better left of as this usually produces
really embarrasing results: aka "Googlish".
I totally agree with the dev's if they say
"translations should be made according to the intent of the word."
I however absolutely disagree with a (weak) statement like
"it is clear that one performs the opposite operation to the other" - that may be so, but it's nonsense nevertheless, sorry, and apparently (Ex-Mambo) Joomla! users are considered total idiots that do not understand "publish" vs. "withdraw" (or any other valid synonym) so they're better confronted with "artificial
unwords" -- despite
"the terms always (?) appear in pairs."
Unfortunately there are
other disadvantages translators have to face. Context, as mentioned above, is among them (the archive/to archive), and it's
unnecessarily the
uncreativity or
unskills of those very translators, or any
unwell-thought "keys" in some language file, but all too often the
unspace "reserved" in the GUI that prevents or at least hinders a
good translation to fit into.
Consequently, it's not excusively "the devs" to be blamed, but screen designers, too: super-tight and precise designs solely focused to english words, like if other languages might not even exist, and zapp: your accurate translation doesn't fit and breaks the design.
Another more important issue, is the sheer ignorance of the
numerus in virtually any "international" application. Even in english. It's so simple to slap a "(s)" to a noun, and everything's fine -- in English.
The translation of an english application is
always a PITA thanks to this really, really -- and I mean
really -- bad habit. Did I mention ignorance?
Maybe it's because too few developers have a background in linguistics. Many know different programming languages but they don't really bother about human languages at all, thus
they don't think (care?) of this very immense problem: to provide an
accurate translation of a sentence or phrase that contains a numerus; accurate by meaning
and grammar.
Translators are doomed to ravish their own language 'cos someone else, sitting in front of his/her english keyboard dictates that plurals are at best anything larger than 1 and a "(s)" is more than enough; the sole solution. And surprise! The absence of something, the nothings, zeros, nulls, may also require a different wording in an unenglish text.
It's been truly surprising that there's been only
one post (by svatas) regarding this issue -- and
it's been ignored throughout. Thanks a lot, svatas, anyway!
I admit, I wasn't aware about the distinction between 1, 2-4 and 5+ "whatevers" in Czech myself. I'm really wondering how some gaelic translator would manage (and suffer) to come around the commonly poor support for
numbering and plurals.
It's not that hard to add support for "numbered words". It can't be accomplished by throwing in a few extra keys, but would require a special syntax to "encode" any of
- no thing
- one thing
- more than one thing
- n-m things
incl.
the order and positiion of number
(s), value
(s), and noun
(s)
I'm pretty confident any
serious translator would accept any such "code" in order to do a good job.
There's no justification to promote crappy language.
Oh, one question: who's in charge for the
quality of the english language files, and the (new) sample content, anyway?
Thanx for reading.
Have fun,
CirTap