From kde-i18n-doc  Fri Jul 23 11:40:40 2004
From: Federico Zenith <zenith () chemeng ! ntnu ! no>
Date: Fri, 23 Jul 2004 11:40:40 +0000
To: kde-i18n-doc
Subject: Re: kolourpaint
Message-Id: <200407231340.43729.zenith () chemeng ! ntnu ! no>
X-MARC-Message: https://marc.info/?l=kde-i18n-doc&m=109058286812060

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Alle 13:09, venerd́ 23 luglio 2004, Adriaan de Groot ha scritto:
> On Fri, 23 Jul 2004, Erik K. Pedersen wrote:
> > There is a string
> > %1 the %2 to %3x%4
>
> And a %1
> and a %2
> and a %1%2 %3%4
> [...]
>             i18n ("<qt><p>%1 the %2 to"
>                     " %3x%4 may take a substantial amount of memory."
>                     " This can reduce system"
>                     " responsiveness and cause other application resource"
>                     " problems.</p>"
>
>                     "<p>Are you sure want to %5 the"
>                     " %6?</p></qt>")
> [...]
>
> So we can substitute in a verb? It's still not pretty.

This is i18n minefield. It is assuming that all languages have a certain 
syntax, and nothing can be farther from the truth.
I would say that fixing this issue is important enough to break the freeze. 
Just imagine a language with the accusative case (say German, Russian, 
Arabic, Esperanto): how are they supposed to know, when they translate 
"image", that it will be used as an object an not a subject?

"Scaling", for instance, can be translated to Italian as "scala", "scalando", 
"scalare", "di scala" (postponed) depending on context. How am I supposed to 
guess? I hope nobody actually expects the translators to go read the source 
code just to figure it out.

I strongly suggest that %1, %2, %5 and %6 in the above example are taken out 
and substituted with separate messages; only %3 and %4 have sense (they are 
numbers).

As a side rant, I would also like to see programmers using more comments, 
_especially_ with such substitutions, but in general with all short messages. 
I recently encountered a "fit label" in Kst that really beat me (is it an 
action that fits a label into something, a label that is fit, a label 
referring to a curve-fitting operation, or what else?), and I had to mail the 
programmer to ask for an explanation (not gotten yet because of time-zone 
lag).

Cheers,
- -Federico
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFBAPk7cbPv7l8OLpsRAoDiAJ0WlgUzKivPiayK7oNyQaX9pvTCzACcDnDR
c7X1xBdY3KFTXo0pkE5pSTk=
=XutK
-----END PGP SIGNATURE-----