At first based on the control’s appearance I just assumed that the Forums translation engine is the same as other SUSE properties, but now that I’ve looked at the page source code, I can see it looks different. Not a sure indication because I can’t see the “code behind” but in general of course all Dev will prefer to simply copy rather than modify without reason.
So, I assume now that the translate control (and backend) is something that is part of the Forums software and not a SUSE/openSUSE add-on.
Here is something then for the Forums Admins to consider… There are a number of other openSUSE/SUSE projects that are using the Weblate translation engine, and are listed next. AFAIK the Landing Page project is the only current web page project, all the others are non-web applications.
https://l10n.opensuse.org/
A question I raised with the current SUSE Weblate translation admin was why each project seemed to use its own individual translation database, that there wasn’t a common database or some automated merge so that translation efforts would be leveraged across many projects, and the answer I received is that each translation stands alone in its context… and that very well may be… I haven’t looked at Weblate long enough to inspect its database format to know whether contexts are included or not.
But, I’d be willing to bet that if any other web property (even these Forums) were merged with at least some of the Weblate translations, it would work… Particularly for what is now translated (Forum labels only). At the very least, fully integrating could mean a <starting point> for translations to be fixed instead of simply being “not available.”
I raise all this, because this is probably a seminal moment for Manager level decision-making, if many projects suddenly have a deep common interest (ie Translation) then IMO it could be an important decision to focus and manage efforts as a single community to leverage each contribution across many projects.
For anyone who is really new to this Translation stuff, although I don’t spend a lot of my time on this today I did some simple investigating about 3 years ago and set up an experimental project of my own using the Google Web Translation Gadget
http://putztzu.github.io/opensuse_translate/
Aside from the idea of becoming closely associated with a Vendor like Google, people should know that this gadget is completely free to use today, you can try out each of the many translations that are possible and notice how well (or not) it works on some technical documentation.
So, technically speaking if someone wanted some level of personally translating even these Forums using the Google Web Translation Gadget, it’s possible (again with the noted association with Google). And, by using the Google Web Translation Gadget, you’re benefiting from years of Google’s translation refinements.
Probably the really big reason to implement machine language translation is for non-web applications. Although there are a few low cost or free Web page translators, it can cost a lot(there’s nothing that’s free or low cost) to provide this functionality for non-web applications and assets.
But, here again this is why I am especially enthused about the idea of possibly integrating these Forums with a common SUSE translation engine, if the individual contributions Users can make to web pages can be leveraged into SUSE UI everywhere, think of the possibilities…
For those of you who are interested in scratching the next layer understanding how machine language translation works, I’d recommend looking up various articles about the Google approach (which seems to be doing a fairly good job after many years of training and data aggregation) compared to other engines which typically lack context support on one extreme or on the other extreme try to parse the sentence into its idiomatic components (I’ve followed a hadoop project that does this and noticed some other similar projects based on Big Data analyzers).
TSU