Last Thursday, I had the pleasure of meeting with Kenn Ejima and Danny Burkes, the creators of Lingr, a unique online chat service built with Ruby on Rails. They were also nice enough to let me share our conversation with you here.

Can you give us some background on Lingr?

Kenn: We started this idea last year. Our parent company, Infoteria, based in Japan, was marketing and selling an EAI software product called Asteria. At the time, Danny and I had come up with plan to move this product to the web, but that strategy was not embraced by our company, so we had to present a real-life example of “Web 2.0” and demonstrate the momentum of web technologies. We went through a ton of brainstorming and prototyping and wanted to build a blog-like app. That was the original idea.

Danny: Then we discovered Comet. At that time, there was no serious “Web 2.0” chat service. So we turned our attention to building a service that could serve as a replacement for commenting on blogs, but in real-time. Eventually, we decided that it would be more exciting to expand that idea, in order to create our own community.

Tell us about your team. How many people are involved with Lingr, and what are their role(s) and backgrounds? And, how long did it take to build Lingr?

Kenn: There are three of us. Danny, Chris Boone (our designer), and myself. Danny and I met through Infoteria. He’s worked with Infoteria in Japan since 1998.

Danny: I have a (very much) varied background. I’ve been involved with software and programming since the mid-80s. Worked for several companies, including Lotus Development, where I met the founding team at Infoteria. I also co-founded a company called iReady, which dealt with embedded software development.

Kenn: I have a long, varied history also. I’ve had a deep interest in computers since the age of 7. I wrote BASIC and Assembler code and built games, and submitted code to magazines at the age of 10. After university, I worked as an Oracle DBA in Japan. I’ve been with Infoteria since 2000. I’ve been in the U.S. only a short time, since last year.

Danny: We discovered Rails in November 2005, and started building Lingr in January 2006. We launched in August. So it took about 8 months, but a lot of that time was our learning curve associated with Ruby and Rails, since it was our first project using those technologies. And, everything we did with Comet was pretty much undiscovered territory.

Tell us more about Comet. Can you describe what it is and how you’ve used it at Lingr?

Danny: Comet is an idea that has many different implementations. The idea is contrary to the normal HTTP open/close request. With Comet, you’re keeping a request open all the time, so when something interesting happens, you know about it right away. And there are various hacks involved to get around the traditional HTTP request/response architecture. The implementation on the client side is pretty straightforward, but the main challenge is in how to support all the open connections on the server side. There’s nothing reliable out there to support this, so we had to invent our own. I’d say Meebo/Writely is the closest thing to what we’re doing in terms of number of connections left open. But no one talks about it. We’ve been very open about how we do it, to share knowledge with the community. It is a very small community - because it is so difficult.

So why Ruby on Rails?

Danny: The best thing the community ever did was the put up the screencasts. I could have wandered into the Rails site and poked around and probably forgotten about it. But the screencasts tell a story, made Rails compelling. I was originally pointed to the Rails site by someone else at Infoteria. And I went there purely out of curiosity. It was only Kenn and I working on this project, so it wasn’t some grand evaluation. It was, “you want to use Rails? Yeah, sure.” We figured if it was broken, we’d try something else. I was most impressed by the momentum of the community and the big wave it has created. And it’s definitely addictive. I think one thing that’s working against Rails in the enterprise is that it’s sold by 37signals alongside a philosophy. So a lot of times selling Rails into the enterprise is more about selling the philosophy than the technology itself. To be more successful, we need to separate the philosophy from the technological advantages.

Any challenges or anything that is particularly exciting that you’ve learned about Ruby on Rails (or related technologies) during this process?

Danny: Too numerous to mention. To me, the core value and most compelling selling point of Rails is ActiveRecord - i.e. no configuration files, automatic ORM mapping, etc. It’s an amazing productivity increase. Having learned the whole platform on this project - the whole thing was discovery. One of the things we tackled was internationalization. Lingr is a completely UTF-8 application. The state of Unicode in Ruby is known - it’s broken, and there’s a lot of work to be done there. But for database-driven apps, where data is stored and presented, databases support Unicode just fine. It’s just a matter of getting the encoding right in the DB, and presenting it correctly in the browser. And, as I mentioned before, almost everything we did with Comet was breaking new ground.

What’s next for Lingr?

Kenn: Our goal is very simple - to make Lingr popular. To do that, we have to support users, and get them to use our product more frequently. We have many, many ideas to forge ahead with Lingr. We’re adding support for client applications to connect with Lingr. We’re starting with our own client applications, and at the same time creating an API for developers to build their own apps on top of Lingr.

Thanks, Kenn and Danny, for the great conversation. And best of luck with Lingr.