I am trying to understand what testtool is doing.
Firstly, we localization projects translate a huge interface file which represents _all_ the strings in the OpenOffice.org interface (the tooltips and Help are referred to as "the Help" section).
This file contains some testtool strings: the strings for the testtool interface itself, and for its error messages.
(Note: the error messages do not display correctly for me, but that may be an OSX-specific problem. All the non-ASCII characters are munged. The testtool interface does display correctly.)
Are the testtool strings in the interface file different from the testtool strings you're discussing in these mails?
It has confused me that testtool has complained about missing localizations. The Vietnamese localization for the interface file is 100%. There should be no missing strings.
On 13/01/2007, at 1:46 AM, André Schnabel wrote:
Thorsten Ziehm schrieb:Hi Andre, as Vito wrote, behind each localization in the TestTool scripts apossible problem should be checked. Only when special action is needed for a special language or a special entry in a listbox runs a special action, then in the TestTool script localized strings are used. I hope only then.Therefore the testscripts have to localized.Yes, that's obvious.But I wonder if the localized strings are really used to test functionality in a localized version or if they are used to check if if a string is localized correctly. To make it (hopefully) more clear what I mean ..Is the localized string for the Writer File Filter ("OpenDocument Text (.odt" in German) used to pick the file filter in the save dialog- or -Is the localized string used to see if the 1st entry in the filter list is correctly translated.In the first case we need the localized string to trigger the correct action. In the second case we do not need the string to test functionality (we just pick the 1st filter and save as writer document).So .. we might be save in functionality testen even without the localized strings and I'd like to remove thoses strings from places where they are not needed (or at least don't want to add more strings if not necessary for testing of functionality).
During our recent l10N QA IRC meeting, the difference between testing "language" and "function" was stressed. Any testing of the quality of the translation is outside the purview of testtool, isn't it? I don't see how testtool can check "if a string is localized correctly". What is it doing?
When we talk about localizing the testtool scripts, it would be a huge help if we didn't have to take on yet another scattered group of strings. Already, we have the templates, samples, formats and autotext strings which don't exist in the interface file. Anything not included in the interface file is going to fall off the priority list for projects with few resources (like mine), or simply be unknown. When I submitted our complete translation for OpenOffice.org 2.1, I didn't know about the templates, samples, formats and autotext needing translation: AFAIK there is no information specifying this in our l10n docs.
If it's viable, I would suggest recommending any further strings be integrated in the interface file. Yes, it makes it larger yet again, but at least we have direct access to the strings, they get updated with the main file, and they can't easily be missed.
But when assigning strings for localization, please bear in mind that only a few languages are supported by Sun. The rest of us are doing all the work voluntarily in bits and pieces of spare time. The current OpenOffice.org translation (interface and help) is half a million words. That's even before you start translating webpages, wiki page, working on spellcheckers and dictionaries, templates, samples, formats, autotext, release notes, install docs and promotional information.
Imagine yourself the only translator for your language.from Clytie (vi-VN, Vietnamese free-software translation team / nhóm Việt hóa phần mềm tự do)
http://groups-beta.google.com/group/vi-VN
PGP.sig
Description: This is a digitally signed message part
