Four weeks ago, we launched a little side project called typewar, which tests how well people know their typefaces. This is the first of a series of blog posts about the site and will explain how the site came about and why we did it. Later posts will go into some of the more interesting technical aspects behind the site.

I only came up with the idea in mid-September and initially did some mockups in my spare time before a couple of days of sprints with the rest of the team to get it implemented. Since then, we've done a couple of sprints to add some new features and there's a lot more coming in the future. Of course, it's all built with Django and Pinax.

The site was announced entirely through Twitter. We created [[|@typewar]] on September 30th then opened up a private alpha the following day that people could join by request on Twitter. We had about 50 alpha testers. The site was launched on October 9th and had over 200,000 page views from 4,800 visitors on the first day. Our only form of promotion was Twitter.

In some respects, the core mechanism of "consider X, now click on either Y or Z" is a recurring pattern in a number of pre-Eldarion sites I've done: [[|Quisition]], Cats or Dogs, Potter Predictions. Even our entry in this year's Django Dash, sortthings, had an element of this.

But behind this simple, widely-applicable mechanism there are a lot of interesting things going on: from how we generate the glyph images, to our rich statistics and use of bayesian averages; from our data denormalization and caching to our use of game concepts like points and levels to make the site more addictive. We'll talk about these in subsequent posts in this series.

So far, around 18,000 people have played and have been asked 1.2 million typeface questions. Over 1,100 people have authenticated (currently via Twitter, but other mechanisms planned) to persist their progress and be able to track their position relative to their friends and others. We average 36 page views per visit and the average time someone spends on the site is 5 minutes.

Why did we build this site? Amidst all our client work and some of our own larger sites that we're working on, why did we take a few days out to build this site? It was actually a fairly calculated move. We wanted to:

  • have fun working together building something together as a team
  • quickly develop components that we know will be useful for sites we plan in the future or are currently working on
  • test out the performance and scalability of some of our implementation approaches
  • study how to get users more engaged in a site
  • promote Eldarion, especially amongst the web design community

Things are going very well so far. Some people think it's too easy, some too hard (we've already been making tweaks) but a huge number of people have said how addictive it is.

We have a lot still to come, both on [[|]] and this blog. So stay tuned! Follow [[|@typewar]] to stay up-to-date with the latest on the site.