Hi Chris,
On 06/08/2012 19:04, Chris Leonard wrote:
Dear Sophie,
First thanks a lot for your feedback and sharing your experience here,
it's much appreciated
Sadly, the L10n community is a very poor place to recruit QA testing.
Why? Because just as the new features are rolling in that need
testing (on a tight release schedule), the new strings are rolling in
that need L10n. There are only so many hours available between
feature freeze/string freeze and release.
Yes, I agree, but l10n people may be in contact with other actors or users
in their own language, this is why I asked here.
That is not to say that all community members do not bear some
responsibility for quality, just that the localizer's busiest period
exactly coincides with the critical testing period towards the end of
a release cycle.
Yes, it's true, however we have nightly builds that are available before
localization cycle. I'm well aware that l10n people usually don't deal only
with our project so I understand it's not easy at all.
The bug in question (spell-checking) is not even directly a L10n
related bug, but it *is* an issue that is very near and dear to the
hearts of localizers, especially those engaged in localizing word
processing tools. Many have spent countless hours, above and beyond
UI L10n, developing word lists and working to enhance their native
language spell-checking capabilities and as noted, many are
more-or-less "teams" of one. This may account for some of the passion
evoked.
and I really understand it, being a localizer in more than one project too.
Automation of testing is an excellent idea, collaborators/allies in
that area may often found among those people who have a focus on
accessibility (a11y), the reason for this is that the alternate
input/output mechanisms introduced for the purpose of a11y are exactly
the sorts of hooks that robotic scripts need to do automated testing.
Personally, I've always viewed a11y as another form of i18n/L10n, but
there is often a small group in any project that is most passionate
about a11y issues and they are very valuable contributors indeed.
This is very interesting, I'm used to test the FR version using Orca, but
I didn't viewed it that way.