Re: String additions to 'anjuta.master'

[F Wolff <friedel translate org za>]

Op Do, 2010-09-23 om 16:40 +0200 skryf Johannes Schmid:
> Hi!
> 
> > It isn't necessary to review things by hand only. I wrote a page on the
> > wiki a while ago about testing i18n.  A big part of this is about
> > spotting strings not marked for translation:
> 
> This is even more manually actually. In a 100 000+ lines applications you
> hardly ever see every dialog. And I am not working in the C locale btw,
> still I never saw these strings because they appear as tooltip somewhere.
> 
> Nothing against the tools but it doesn't improve that situation that much
> unless you have time for a full test of an application which is
> impossible.


Hi Johannes

As someone who as done this, I think you are overestimating the work,
and underestimating the value.

If a developer is testing some new feature, and cares about i18n, then
it is easy to just test the i18n of the new feature that is being worked
on.  If a developer on the team bothered to write a tooltip, I guess
that same person might care what the tooltip looks like, so might review
it in the application anyway.  So if testing this is hard, testing the
application is hard.  I don't think testing the i18n is really that much
of a burden.  The question is if developers want to test, and if they
care about i18n.  I guess that most care.

Podebug makes it really easy to make a pseudo-localised file to test if
a 100% translated file gives a 100% translated GUI.  It can also help to
test a few other aspects of i18n (like handling non-ascii text,
right-to-left text), and I think is a way to make international user
experience better from the start while the code is fresh in a
developer's mind, rather than waiting for bug reports necessitating
rework in the code after the fact.

Regards
Friedel

--
Recently on my blog:
http://translate.org.za/blogs/friedel/en/content/pootle-210-released