Forum Moderators: Robert Charlton & goodroi
1) Will I get in duplicate content trouble?
2) Should I go for en.example.com/ or www.example.com/en/? (I would prefer the subdomain approach)
[edited by: tedster at 5:21 pm (utc) on Sep. 12, 2006]
[edit reason] use example.com [/edit]
I would prefer the subdomain approach
Many webmasters would -- but my experience is that using directories is much less problematic.
1. The average person does not understand subdomains. They will still type in "www.subdomain.example.com". Yes, you can and should make that 301 to subdomain.example.com to handle crazy IBLs.
2. Directories are seen as part of the principle domain. Subdomains are treated a lot like an independent new domain and need to grow their "trust" factors more intently.
It doesn't matter if a German page, for example, gets filtered out of the English language results. But since you have a lot of English words even on the translated pages, you do have a challenge to make sure the German page comes up on the German results. You want the English page to be filtered out there
So make sure you are giving out as many unambiguous language signals as you can. Especially, make sure the server headers declare the correct language -- rather than hoping the algo can sort out the correct language for the mixed words on the page. Also use the lang attribute it the html tag -- <html lang="de">.
Finally, I suggest you notice and respond to the actual search results for each language, rather than being too concerned about what you see on a site: search.
They have both a domain blah.de and de.anotherblah.com
de.anotherblah.com shows up in the global results of google.de
but if you try a local search you can only find blah.de.
The IP's of both de.anotherblah.com and blah.de are in the US.
However, this thread was specifically asking about duplicate content filitering - I apologize for my diversion of the topic. I would also add if there are pages that are supposed to be non-English but the content is still predominantly English technical terms, then perhaps that English content could be inserted via an iframe, to keep the major part of the html document non-English.
