[Kst] Fwd: Re: translations

[George Staikos]

Kst developers, please read this and keep it in mind while writing code.  We 
should probably audit our i18n()s soon to fix any existing problems.

----------  Forwarded Message  ----------

Subject: Re: translations
Date: Saturday 07 August 2004 16:06
From: Federico Zenith <zenith at chemeng.ntnu.no>
To: George Staikos <staikos at kde.org>

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

Alle 22:11, giovedì 05 agosto 2004, hai scritto:
> Hi Federico,
>
>    Thanks for the great work on the Kst translations.  Is there any chance
> you could compile a list of all the confusing or improper i18n()s we have
> so that we can go through and fix or clarify them?

Hi, I went through Kst's .po file to look for possible improvements. Overall,
Kst is definitely not a "bad" file, but I'll address some general issues and
some specific examples.

In a nutshell:
- - Comments are needed when using technical terms and/or short messages.

In detail:
- - When a message is in a certain scope, it should be declared with a short
comment, like "Plotting", "Interpolation", "Plugin", "Options" or anything
that can reasonably give a sense of context. This is true especially for
technical terms: e.g. "Covariance" has normally a unique translation in most
languages, but if the translator did not study the English translations of
statistical terms, a bad/misleading/wrong translation might follow. A simple
comment as "Statistics" will help a lot.

- - Also "FFT" would probably need an explanation - I knew it was a Fast
 Fourier Transform, but that's just luck. Other languages (I would guess
 French) might have different acronymes.

- - A comment is unnecessary if the message is sufficiently long to be clear
 by itself; on the other hand, the shorter the message, the more necessary a
 comment will probably be.

- - English has some words that are wildcards. "Fit", "Hit", "Put", can have
several different meanings (beside almost all verbal forms). In general, most
languages are fussier than English, and differentiate between them. The
string I mentioned some time ago ("fit label") is a perfect example. In
theory message 109 too could be confusing: "curve name" could be interpreted
as "bend the name" (as one can do, say, in photoshop), even though this is
definitely more far-fetched.

- - What's the difference between arrays and vectors anyway? In Italian they
 are usually both translated as "vettore", so having them both is a bit of a
 problem.

- - The "fit" word needs some scoping. I'm not always sure I'm working on a
 data interpolation or on the autofitting of plot elements.

- - "Re-grid" is a bit unclear. I guessed it is a redrawing of the grid
following, maybe, some plotting operations. I'm not totally sure though, a
comment would fit in message 699 (rc.cpp:1030)

- - When using simple translations containing incomplete sentences or single
words, give what they are referring to. Most languages (English being a
notable exception) do have genders for words: an "interpolation" and a
"label" are ladies in Italian, while a "plot" is male. I know this is insane,
but that's how it works :-). Number is also important, for some languages
other factors too (e.g. determination in Scandinavian languages and German).
So, give a short example of the sentence in practical use; see message 720
(rc.cpp:1083) where "logarithmic" is left alone. And yes, a "logarithmic
scale" is a lady in Italian, but a "logarithmic axis" in a dude.

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

iD8DBQFBFTZAcbPv7l8OLpsRAuxkAJ9shn1ofmpfmkIq6ac7rydIubXTQwCbBZCK
WQLhmV8904bmqjm4AiMagPw=
=cy3Y
-----END PGP SIGNATURE-----

-------------------------------------------------------

-- 
George Staikos
KDE Developer				http://www.kde.org/
Staikos Computing Services Inc.		http://www.staikos.net/