Tomomi and I and our friend Taku Nakajima have been working on a project for the last year and a half called Cojiro, which we submitted to the innovation awards last week. The project is to build a web platform for what we call “cross-lingual curation”: curating content from one language to another. We’ve actually already built a prototype (see below) which we’re currently testing — if we get the grant then it will fund the next stage of the project, which will be entirely open (for more info see the github page we just setup).
Of the proposals so far, the closest in spirit to what we are doing is Paula’s on social news gathering. Cojiro actually began as an idea for a multilingual newsroom for GV Japan. We used a bug-tracking tool called Redmine which Taku setup for us to post story ideas and discuss them (internally). We still use that system, but there are a lot of problems with it, and other popular services (Storify, etc.) and tools (wiki, blogs, etc.) didn’t really do what we wanted either.
Here are some of the things we’ve observed from our experiences:
- Popular social media and curation services are not very good at handling more than one language at a time. If you upload a picture to Flickr, or a video to YouTube, or tweet something, there is no easy way for someone else to come and add a translation. There are ways to get around this, for example the way Meedan turned curated.by into a curation/translation platform, but this feels like a hack — there must be a better way to do it.
- Translation platforms are not the solution either, because they are made for translating and not for curating/news-gathering. Translation is a part of the process, but it is not the only thing or even the main thing.
- Even when we do want to translate, we can’t translate everything — there is just too much of it! So, we want to focus on the important parts. The caption of a popular video or photo is very important. A key quote in an article is very important. The title of a blog post is important. On the other hand, one line of text buried in a long article is not as important.
So, the key thing here is figuring out which information is valuable to translate, and making it really really easy to translate that information. That’s what we’re trying to do with Cojiro.
Another thing that’s worth mentioning about Cojiro — similar to what Paula proposed — is that it’s completely open: GV can use it but so can anyone else. If this works out, GV language communities could use this as a way to reach out to readers and other curators to get them involved in the news-gathering/storywriting process. Paula’s example from the Guardian shows how powerful this kind of collaboration can be.
If you’re interested in this idea please let us know in the comments! And for more info have a look at the links below:
- Prototype that we’ve built and are currently testing. This is setup in Japanese/English, but the system can be configured for other language pairs.
- An example of a cojiro thread: Co-working places in Tokyo which became a GV article: Coworking Spaces and Nomad Workers in Japan
- github page for the new codebase (scroll down for a summary) — we’re building the platform in Ruby on Rails.
- And just for your enjoyment, a picture of us learning to code (taken a year ago) — lots of work but also lots of fun! 🙂
Edit (2/2/2012): Added screenshot.