I hope Duolingo will explain the standards it uses for material that is fit for uploading. A lot of my attempts to upload stuff have been rejected with the message that they "may have offensive content." The most recent ones have been recipes. Even Italian recipes aren't that sexy. What's the problem?
If the recipe contains a "handful of (chicken) breast" then you're going to have difficulties.
Frankly, I'd prefer it if they just had a mature content flag, and still allowed the material. Sticking entirely family-friendly is kinda an unrealistic expectation for learning a new language.
We use a simple word-based classifier, and it's certainly not perfect! If you post examples of inoffensive documents that were rejected, we may be able to improve our classifier to let them through.
Thanks, but please let me understand. Does the classifier process each language separately, or does it treat all the words as English words?
Then the Italian one has a problem. Could I possibly see the list or whatever it is for Italian? If you don't want to post it (Evidently you don't) could you send me an e-mail? I'd like to know what to avoid in the future.
Our goal is that you shouldn't have to avoid anything that's not really offensive. I think it's best if you send examples of links that are being rejected when they shouldn't be, so that we can use those examples to improve our classifier.
Looking at the Cesare Borgia page, it does contain the words "naked body" (in the English translation - I can't read Italian).
EDIT: and the recipes contain "extra virgin" (olive oil)
Out of curiosity, how does your filter handle humorous misspellings? We inexperienced Italian students tend to drop double consonants, and doing so for "penne" might get us in serious trouble.
Here are a few recent examples. I hope you'll let me know what I should avoid in the future. http://it.wikipedia.org/wiki/Cesare_Borgia http://www.ricetteitaliane.info/sicilia/involtini-di-pesce-spada.html http://www.ricetteitaliane.info/sicilia/polpettone-alla-siciliana.html http://cultura.panorama.it/serie-tv/commissario-montalbano-serie-rai http://www.repubblica.it/tecnologia/2013/07/09/news/stampanti_3d-62671582/
I've updated our classifier, and the links you posted should work now. Feel free to send other examples if you continue to have problems.
This is what I love about Duolingo. "Oh, you're having an issue? BAM, not no more!"
I assume you mean "check for certain words" with "simple word-based classifier". Couldn't you use a run of the mill greedy Bayesian system to filter content? That way you can cathegorize spam as well, and preevaluate the perceived difficulty of a text. Also, you could run whatevery training sets you want to feed it with during less taxing periods of time.
I, for one, am happy that Duolingo leans towards over strictness when determining offensive content. I am comfortable recommending Duo to my teenage granddaughters with no worries. They love it and use it all the time! I think everyone would agree that it's a good thing when teens actually want to learn and experiment with different languages.
I think it is good that the site is family-friendly too, since it has happened that I recommended it to a couple of youngsters and very religious adults.
But maybe some people are fine with (and want to learn) things that might be considered mature, including news stories and song lyrics. Perhaps the family-friendliness could be maintained by asking the user's age during sign-up (and only allow mature content to be viewed by people who have given an age, and that age should of course be 18+, and perhaps even still ask people to opt in, since I understand that there are grown ups who do not want to review mature content). :)
I like your idea. A type of parental controls. I'm just the grandmother, so, of course I always ask my daughter if she is comfortable with sites I recommend to my granddaughters:) My point was that we should be encouraging kids to learn languages. DL's game like interface is what encourages them. Most of the language learning sites for kids are, well, I guess a 16 yr old would call ''babyish'' and she wouldn't be caught dead using one of those sites. She's a teenager:)
I like the idea of parental controls. I don't like the idea of withholding important knowledge from adults.
The problem is that all languages have "Slang" meanings for thousands of otherwise innocent words.
For example, I have a friend from Quebec who informed us that they (in French) use the word for "cat" in a way I'm sure duolingo would never have warned me of. We have something like that in English too and I certainly wouldn't want to be throwing the word around in ignorance!
How does the comment box know it's a recipe? Of course, I agree with you that recipes are harmless and actually add cultural aspects to learning not only a language but the context in which it is used. Understanding context is crucial to understanding a language and how it is used.
It's obviously a fault on their part. Though there's not much you can do about it. But I agree with fourohbro, it'd be much better to just introduce a mature content flag. Not everyone here is twelve, obviously, and might have different questions that may be deemed unfit for young children.
I just had this link get refused - http://www.eltiempo.com/politica/buque-estadounidense-estaba-sin-permiso-en-aguas-colombianas_13170080-4.
What's annoying is that there's no easy way to report these - it would be good if there was a link on the submission page, the same as are on the lesson pages.