I'm running an internationalized web service with user-created content (study notes). Most content is from Polish version, although we have younger Turkish and Russian versions too. Each version uses its own domain (let's say it's
russianversion.com) and there is a language toggle available to switch the language of the interface (header, footer, menus), while of course content remains untranslated.
This leads to repeated content - a given Polish study note exists on all 3 domains:
polishversion.pl/note/123 - Polish interface + Polish content
turkishversion.com/note/123 - Turkish interface + Polish content
russianversion.com/note/123 - Russian interface + Polish content
My first instinct was to add "noindex, nofollow" to the Turkish and Russian version of a Polish note, since in most cases it would be irrelevant for these audiences (but not always - probably study notes for a French as a foreign language course are relevant to all).
Then I came accross
link rel="alternative" hreflang. According to this article: http://googlewebmastercentral.blogspot.com/2010/09/unifying-content-under-multilingual.html I should be fine adding:
<link rel="alternative" hreflang="tr" href="turkishversion.com/note/123" /> <link rel="alternative" hreflang="ru" href="russianversion.com/note/123" /> <link rel="alternative" hreflang="pl" href="polishversion.pl/note/123" />
to the Polish version of this note and:
<link rel="canonical" href="polishversion.pl/note/123" />
to both Turkish and Russian.
I'm still having doubts though, whether this won't harm my Turkish and Russian version which have very little of their own content content yet. So the whole issue reduces to:
Is it better for the Turkish version's seo ranking to have:
100 000 pages indexed of which 99% are Polish content + Turkish interface (with a canonical link to Polish/Polish) and only 1% are Turkish content + Turkish interface?
Only 1000 pages with Turkish content + Turkish interface indexed (and Polish notes never indexed with Turkish menu at all)?