Startup Weekend

We launched Scrolltalk last weekend, as a part of Startup Weekend West Lafayette. It was a hell of a ride, and we were the first Startup Weekend to launch our product by Sunday. We skated into the deadline, going live at 11:59:57 on Sunday night - we were ready at around 1:30pm, but testing a few servers forced us to upgrade, and upgrading our server configurations took a while.

So, how did it go? We came in with no idea what we would be doing - unlike other startup weekends, who debated for a week or longer on the blog before the weekend, none of us really had any idea what to expect. We spent the first few hours brainstorming - we got a bunch of ideas, ruled out a lot of them as unfeasible over the course of the weekend, and narrowed it down to four. After enough debate, we didn't like any of those ideas, so we scrapped 'em and started over. Then we hit upon scrolltalk - an evolutionary chatroom that watched what you wanted to talk about, and brought you together with other people who wanted to talk about the same things. No rooms, no user lists, just you and the people you want to talk to.

We bought a domain name, started getting ideas and planning the program architecture, and called it a night around midnight.

Saturday was a wild ride. We had a lot of debates about what architecture to use, what platform to go with, what the structure of the app should be... But by Saturday evening we had an "ugly duckling" version of the chat going. We had it working with tags and filtering conversations and doing everything it needed to. We had some more debates, talked about the company, worked on PR stuff, and eventually called it a night around 12:30.

Sunday was pretty crazy as well. What were just going to be a few tweaks on Sunday morning turned into a bunch more features we could add by launch time. I was wary about how many toys and niceties we could realistically add without breaking the code in some way, and still be able to launch on Sunday. But between the devs working on the code, we managed to get a lot done, and added a whole bunch of features to the app. We did a "Delta" launch (think beyond beta, man!) to our email list subscribers, found some server problems, and closed down to work on them.

By Sunday night we had switched over to the new servers, and after some last-minute changes, we were ago! We were the first Startup Weekend event to launch our product by Sunday.

So what did I do in all this? I wrote code. Here's my contribution in about 20-24 hours of coding time.

• Wrote out AJAX with Prototype to poll our chat service and update the client's window.
• Wrote out php to poll the chat db, filter down to user conversations which were relevant to the client, and display them with proper markup so our designer could work with them.
• Wrote out AJAX and PHP snippets to deal with adding tags and setting users you always want to follow.
• Wrote regular expression filters to watch for markup in the chat message and apply it (try typing +tags into the chat box!).

It was a huge amount of development, and if it weren't for some other really talented folks, Scrolltalk certainly wouldn't have happened. Thanks to the rest of the team for all your contributions - we made something awesome this weekend, and I feel very proud of our baby.

As for where scrolltalk is going -- well, we're an LLC now, and we're looking at what it will take to make scrolltalk the most interesting way to connect with people on the web. We'll be working on it, and in touch.