Apply HN: paralleltext.io – Learn languages by reading

[mstipetic]

http://paralleltext.io enables a simple way of learning languages by reading and listening human translated books in parallel.

Problem: Learning languages is a difficult process that requires hard work no matter if you choose the academic way or to use existing apps. It also requires continuity, a key factor which is hard to achieve with the available methods.

Solution: We want to change that by combining the learning experience with the pleasure of reading and listening to a book. A large number of book translations is available, but we think they're being underutilized for this purpose.

Progress: We built a tool that matches human translated paragraphs from ebooks in all European languages and a novel interface to read them. Now we are focusing on adding more books and more engaging features.

Market: Language learning and publishing

Team: We are two developers, both with product management experience living in Berlin.

Feedback: we got amazing feedback so far (https://www.reddit.com/r/InternetIsBeautiful/comments/3x2tux/paralleltextio_improve_your_knowledge_of_foreign/)

[xiaopingguo]

I'm hugely in favor of graded sentence level translations especially for adult language learners. It is a more natural form of spaced repetition and requires much less will power to stick with it. Finding decent suitable corpora is the difficult part but very amenable to crowd sourcing efforts.

Also see this link for why reading is a really good way to acquire language: http://llt.msu.edu/vol12num1/mcquillan/default.html

[mstipetic → xiaopingguo]

Sorry, what do you mean by "graded sentence level translations?"

Also, this article is really supportive of our project, thanks!

[xiaopingguo → mstipetic]

Sentence level translations, rather than word or paragraph level, graded by difficulty as beginner/intermediate/advanced.

You can look up more about that research from here: https://en.m.wikipedia.org/wiki/Extensive_reading

[mstipetic → xiaopingguo]

Ah, ok, now I understand. The problem might be that we aren't using machine translated texts, but human translated. So when matching translations, often sentences don't have a 1-to-1 matching. Translators often change sentence structure to better suite the new language (there are even instances of translators adding new characters in the book).

In any case, it's a very interesting problem

[tgflynn]

I think it would be great if there were a way to mouse-over a word and get a popup with the word's translation and information like part of speech, conjugations, declensions, etc.

[mstipetic → tgflynn]

Yes, we're working on it now. We have to both make it seamlessly integrated in the interface (a popup might be too intrusive), and unfortunately the state of open source dictionaries is not ideal (which we plan to change also).

Cheers!

[kumarski]

The books you're offering up have a ridiculously difficult readability. How do you plan to change that?

What languages will you target first and why?

I'm quadralingual as well and blog about language stuff once a year or so. Holler at me if you want poignant feedback. I've been published on lifehacker and other sites for my language learning tips.

[mstipetic → kumarski]

Hey, thanks for the feedback!

We've been so far working with books that already had some translation metadata which we could use, but just recently we got a tool working which does it automatically. It's still a work in progress, but we managed to put a new book out, Siddartha by Herman Hesse in English-German combination.

We're gonna concentrate on English-German combination first, since we're in Berlin now, and can calibrate the matching tool ourselves, but adding new books in new languages should more or less become trivial soon.

We have a lot of features in the works, one of them being providing and recommending books with different levels of difficulty.

Any help or feedback would be very welcome, you can always reach us at contact@paralleltext.io

[kumarski → mstipetic]

I highly recommend using popular short stories.

Books are far too long.

Short stories give the user a sense of success. Even just covering news articles would be pretty solid. :)

Current events provide a lot of context.

[mstipetic → kumarski]

Thanks!

We might look into short stories also. News articles might be interesting also, but sourcing multilingual new articles might be a problem. In any case, we have a lot of work in front of us :)

[dputtick]

This is a cool idea, and something I would consider using. Two questions:

- How are you sourcing your translations, and what is your technology solution for matching them up?

- Do you have plans to make a native app for tablets that can store the books locally, or are you going to move ahead with improving your web tool for now?

[mstipetic → dputtick]

Thanks, we think it's cool too!

The new books we're adding come from Project Gutenberg. We parse the ebooks, and using our tool and a combination of open source technologies (http://mokk.bme.hu/en/resources/hunalign/ for instance) we do the matching and then manually review the result, if necessary.

We started building an app, but then realised that the interface would basically be the same, and a lot can be accomplished with modern web technologies. We plan to leverage local storage and to make it feel as native as possible for now, since we have fairly limited resources and a lot of work in front of us.

[dome82]

I really like the idea!

It would be nice to have features like the one in ReadLang where I can click on multiple words and get them translated. I like the part where I can listen the sentences.

Are you going to apply this to any article on the web? What will be the business model of your service?

Good luck!!

(P.S: In Norway, we have a similar site that translate the news. Here it is an example and hope it can be useful to you: http://www.klartale.no/kultur/kunne-aldri-v%C3%A6rt-forfatte...)

[mstipetic → dome82]

Thanks!

The thing is, these aren't automatically translated books, we're matching real book translations, so at least our approach isn't applicable to articles online.

We have a lot of features planned, from word lookup, to hooking them up to real audiobooks, and a lot more.

Our motto is "Provide maximum good" so we don't want premium-only features, and all the books in the public domain would always be free, but as our business model, we'd like to offer recent, in copyright books in this format for sale.

[astrange]

This is a good idea. What makes it a startup and not just an app? What would you reinvest the revenue in?

BTW, the only book posted for English to Swedish has a pretty literary translation and long sentences, so it's hard to learn anything from it.

[mstipetic → astrange]

Thanks!

Well our business model would be recent, in-copyright books in this format. For instance, reading Harry Potter in this interface is something we could imagine being useful. Other than that, we think this tool might be useful in the education system, we can provide a lot of added value there.

Some language combinations are unfortunately underrepresented, but we've just built a tool that matches translations automatically, so that should change soon. Currently we are focusing on English-German combination since that's what we can check manually while we work out the kinks in the tool, and then we will spread out to other languages.

[orangetabbycat]

This is so great.

I wish there was a way to translate individual words rather than entire paragraphs.

[mstipetic → orangetabbycat]

Thanks!

We're working on adding individual word lookup right now, we should be rolling it out soon.