Localization issues

Sun Mar 18 19:40:05 UTC 2018

Hi,

2018-03-18 15:57 GMT+01:00 Michael Bauer <fios at akerbeltz.org>:

> Hi folks
>
> Ok, first of all, we brought gd back up over 95% again. Sorry it's been a
> while but losing *that* many translations in the move to KDE was a real
> kick in the guts.

Great news!
We have a script to automatically fill the translations from the Gtk+
version. We also reworked a lot of strings so they couldn't be reused.

> And now I've forgotten how to commit on the new system, KDE does my head
> in. Would it be possible for an admin to grab the po from
> www.akerbeltz.org/sealach/gcompris_gd.zip and commit it please? :)
>
>
I've commited it on stable branch.
svn co svn+ssh://svn@svn.kde.org/home/kde/branches/stable/l10n-kf5/gd/ to
get the stable branch for gd.
Then you put the file on messages/extragear-edu/gcompris_qt.po and to
commit: svn ci -m "your commit message"
svn up, to update your local files with the one in the repository.

> I've also run into various issues while bringing the localization back up
> which I feel need adressing. In no particular order (though I deem 1, 6 and
> 7 most critical) these are:
>
> 1) the long blocks of text are really unhelpful for localizers. If you're
> working on a laptops and in Virtall (for example), stuff just disappears
> off the bottom of the page in these really long blocks. Working on my
> desktop and in Trados Studios, they sometimes make even Trados crash due to
> sheer length.
> The additonal problem is that when devs change some tiny thing in en-US,
> suddenly your translation ends up with massive chunks of "untranslated"
> text just because ":-)" was deleted from a dozen "strings". Someone should
> really look at breaking those into smaller chunks, certainly between
> paragraphs of text and in future, possible even down to sentences (which
> would also make the use of translation memories more efficience if
> sentences are moved around or split etc).
>
>
Yes, that's one of the problem we have. Strings are mostly made by
developers where they should be done/reviewed by teachers/native English
people to simplify them.

> 2) Columbus Day - is that really such a great example to use, given how
> controversial it is?
>
>
There was already questions about this one (https://mail.kde.org/pipermai
l/gcompris-devel/2018-March/005392.html), we'll remove it and change it to
something more "conventional".

> 3) Mothers Day - this is not on the same date in every country but the
> activity does not seem to take that into account.
>
>
I guess we took the American one without specifying it. But yes, we
shouldn't use celebrations that are not the same everywhere.

> 4) The family activity is very Anglo-centric. Given the complexity of
> kinship systems, this should really be re-visited (check out
> https://en.wikipedia.org/wiki/Kinship_terminology) and at the very least
> be described as Anglo-European kinship model in the activity, at least to
> set expectations.
>
>
We got a comment from the Chinese translator too regarding this. We'll add
that it's the system of most European and America countries and add
contexts for translators to know who we are positioning from.

> 5) Pounds and ounces are not entirely restricted to the US.
>

> 6) The quality of the English is seriously poor in places. Given that
> GCompris charges money for the full version, I strongly feel that before
> more activities are addes, the existing ones need proofreading by native
> speakers.

Yes, for sure, this is a place where we lack experience and need to improve
the quality. Most of the texts are written by non native English and we
don't have native people reviewing the texts. Do you know people that would
be interesting to help reviewing the activities?

To give a few of the worst examples I've seen:
> This activity will give teach about who we are related to our relatives
> (I think this means "This activity teaches about how we are related to our
> relatives")
> For the married couple we have highlighted with the rings that is between
> two circles. For the rest of relations we have only solid lines.
> (I think this is supposed to mean something like "Married couples are
> marked with a ring linking two circles. All other types of relationship are
> linked with a solid line")
> Red circle will point to you and blue one to your relative. Now you have
> to find what you should call this person.
> (I think this is supposed to mean "The red circle marks you and the blue
> one your relative. You now have to find out what you call this person")
> Now he wants to give 4 candies to his friends.
> (in English, candy is "pieces of candy", you can't use "candies")
> You can't put more than  candies in the same widget
> (is this really an appropriate in-activiy string, "in the same widget"?
> That's suddenly very technical for an activity where someone is sharing 8
> pieces of candy between three people...)
>
> 7) The Monuments activity is nice but needs a content review. Apart from
> the fact it seems very India-heavy, there are large sections with
> questionable English and content, which in parts seem very technical, in
> parts like advertising.

There are several levels, a global one with the 7 new wonders, and the
other ones corresponding to other countries.
Regarding the content, I think it was mostly inspired by wikipedia. But as
you say, it can be improved and simplified for children to only describe
the monument.

For example:
> To pen-down the philosophy, ideology, the inner and outer beauty, as well
> as the historical legacy of Sri Harimandir Sahib is a momentous task. It is
> a matter of experience rather than a of description.
> (that may be the case but in a description for a monument seems out of
> place)
> There are 30 caves in Ajanta of which 9, 10, 19, 26 and 29 are
> chaitya-grihas and the rest are monasteries.
> (the last part is very technical (do we really need the cave numbers
> rather than saying "5 of them"?) and chaitya-griha may make sense to Hindi
> speaking children but anywhere else, without an explanation, this is rather
> pointless.
>
> Over and out
>
> Michael
>

Thank you for reporting the issues, it is always interesting to have
feedback to know where we should concentrate more.

Johnny
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/gcompris-devel/attachments/20180318/d8f20b42/attachment.html>