Suggestion: Allow learners to translate any web page but limit distribution for copyrighted pages until owners pay.
(By "any web page" I mean any page that passes Duolingo's filter.)
Moreover, allow site owners to see how much translation is going on for their site. If there is a lot and the pages are copyrighted, then they may consider paying to make the translations publicly available.
The thing with copyrighted pages, though, is that we are violating the copyright by translating them without permission, which could cause problems for Duo. Translation (even if not-for-profit) is not generally considered to be "fair use" as it is considered to produce a "derivative work". (As far as I understand it... I did some research on this topic a few weeks ago, but I am not a lawyer or anything). Also, all the translated articles on Duo are "publicly available"... they're here, they're findable via search engines, they can be re-distributed throughout the internet.
It can be hard to find things that aren't copyrighted or are in creative commons, but it is not impossible! If there is a topic you are particularly interested in, try doing a search for something like "articles on (topic), creative commons" in whatever language you're looking for articles in.
Google translate works on copyrighted pages. Is that a violation of copyright law?
As I said, I'm not a lawyer or anything. With Google translate, it's a machine rather than a human translation (which may or may not affect anything) and also I don't think it stores or distributes a copy of the translation. All the articles on Duo are stored and anyone can access them. A Google translate webpage translation disappears.
Well I did say that a translation would have limited distribution until the owner pays.
P.S. What about Google caching pages? Is that a copyright violation?
I guess Duo could limit the distribution of those articles specifically, but as of the moment all translated articles are available to everyone on the internet, which is what I was basing my comments on.
I have no idea if Google caching pages is a copyright violation.
So what happens to the idea of translating the internet? How much is not under copyright?
I don't know the percentages of copyrighted/not copyrighted. But we can translate copyrighted articles, we just need to have permission to do so.
Who's going to negotiate that with a copyright holder in another country who may not speak English? What's the incentive for that person to agree?
Someone who speaks that person's language. A large percentage of Duo users are people learning English, who speak Italian, Spanish, Portuguese, etc. as their native language.
I imagine in most cases there wouldn't really need to be negotiating. Simply saying "I'm learning Italian and would like to practice by translating your articles, is that okay?" would probably get a positive response. The opportunity to spread brand awareness to people in other countries who speak other languages could be an incentive, if they need one.
Have you actually tried to do that? A large percentage of Duo users are people learning a foreign language who don't feel ready to write a letter in it so that they can ask permission to translate something into English. A lot of web sites don't list a person to contact. Besides, you're not just translating their articles; you're putting them up on a web site. Some sites aren't selling brand awareness, though that would explain why I see so many consumer electronics ads in the Immersion section.