I just received this message from Brian McConnell. Brian is an O’Reilly author and blogger, and generally a smart guy. I didn’t know he was working in Rails–I’ve always thought of him as a Pythonista. Anyway, if you’re interested in this project, let him know. Description follows…

> I am writing to invite you to join, and hopefully contribute to, a
> project that has the potential to revolutionize communication.
> That’s a tall claim, but this is a simple idea that solves a tough
> problem. I have been working on this problem for many years, since
> 1998 including early work, and have developed a system that will
> enable publishers to marshall their own readers to translate their
> websites, blogs or any other texts into many languages. Let me
> first describe how this works.
>
> The prototype system we have developed (in RoR if you’re curious)
> monitors a participating website’s RSS feed for new articles. When
> it finds new content, it runs it through machine translators to
> obtain rough, or gisted, translations to about ten languages. It
> creates a wiki page for each translation. The publisher encourages
> his/her readers to go the wiki to edit and improve the
> translations, or start translations to any other language. The
> translations are output via RSS and static HTML. The key insight I
> had was to realize that any website with a decent sized audience is
> bound to have bilingual readers without even knowing it, some of
> whom will be happy to translate posts to their respective
> languages. This will also have a network effect because as a site
> becomes visible in other languages, it will draw in more readers,
> many of whom will speak other languages. It’s a simple technique
> but it could have far reaching uses as the language barrier is
> really the last frontier left in communication.
>
> Over the holidays, I thought seriously about developing this as a
> business, but decided to give this away. This is an example of a
> problem where open source is a superior solution. We have developed
> a pretty good prototype in Ruby, as well as detailed specs that
> describe how the system operates. The problem is that human
> language is tricky, what works well for European languages might
> not work so well for others. I decided to put this out there as a
> reference design and starting point, so that others can take the
> concept and adapt it to their needs, whether by embedding this
> within a content management or blogging system, or by adapting it
> to deal with other languages. The only centralized element of the
> system is a statistics interface that keeps track of the documents
> processed system wide (for counting and search engine optimization).
>
> I am putting this invite out via the O’Reilly lists because many of
> you are developers, and because many of you are bilingual. I am
> looking for people to contribute to the project as developers or
> translators, as well as evangelists (to get the idea in front of
> popular content management systems). My goal for this project is
> ambitious. I want this to be a ubiquitous feature, as worldwide as
> the worldwide web, even if users are not aware of the project
> itself (from their viewpoint, they’ll just see “[Translations]”
> links wherever they go).
>
> See the project’s website
> for more information, including some detailed specs for our hosted
> service. The test system will be live soon, at which time we’ll
> also publish our source so that developers can adapt it to their
> needs.
>
> If you’d like to learn more, or have specific questions, I can be
> reached at brian ||AT|| mcconnell ||DOT|| net.