Of all the mobile games I have tested and played over the years, only one really sticks in my mind -- Flying Toasters. Graying techies already know what I am talking about. Flying Toasters were the ubiquitous screensavers from software developer After Dark. Winged retro-style toasters soared randomly across early color screens to prevent the dreaded “burn-in” from static screen images. It is amazing what passes for entertainment when a technology is young.  A few years ago a game company turned the familiar After Dark brand into a one-button phone game that was the smartest, simplest mobile interface I have ever seen. All you had to do was keep the toaster aloft and navigate barriers by pressing a single button on the phone.  It was captivating -- and a lesson in the need for ultimate simplicity on this platform.

I was reminded of this game by Crush or Flush, a new but very popular social networking, hook-up service. Rather than browse Web sites and publisher pages, users simply browse one another. You can easily enter a photo of yourself and about eight tags that profile you. According to Josh Levine, director of marketing for developer IceBreaker, the most popular tags are “movies,” “Being awesome” and “Booty Shaking.” Yup, Crush is favored by the 18- to 24-year-old demo. But like a lot of mobile-only brands, it is taking off very quickly via viral distribution. In its first 90 days, Crush or Flush signed up 100,000 members; when I spoke to Levine a couple of weeks ago, the total was nearing 150,000. Echoing what other mobile developers are telling me about the college-age market, almost all members prefer to sign up via mobile rather than the optional Web site.

At heart, this is speed dating by phone, but all done anonymously. You can apply filters to the profiles you will see on your phone and at each image and tag set you simply “flush” (reject and move on) or “crush” (flag the person as someone interesting). The system only hooks up two mutual crushes, and even then the chat is anonymized behind usernames. No emails or phone numbers are exchanged.

While Icebreaker has used various forms of Web and mobile advertising, Levine says that much of the distribution is coming virally and because of the application’s very shallow learning curve. It all comes down to three key press choices: 1 – crush, 2 – flush, 3 – tell a friend. Actually part of the genius of the viral element here is that it lets you play matchmaker and pass profiles on to others.

Levine says that the average session length is 20 minutes and the system has generated million of votes thus far, with thousands of mutual crushes. Offering users such a brain-simple, binary choice seems to be essential to the service’s success.

The challenge for marketing and advertising will be how to slip into a content well like this. Levine and company are pondering this now. “People are tired of banner ad nonsense,” says Levine. “We want something unique.” And so they are looking for something less intrusive and contextually relevant. The most obvious thing to do is crush or flush ads themselves. Just slip a product profile in with the personal profiles and let users elect to see more or move on. This would give advertisers exposure, yet remain fully integrated with the playful nature of the content itself. Users already know what happens when they click into a profile for more information, so there is none of the usual inhibiting mystery attached to clicking on a mobile banner ad.

For all we know, Crush or Flush will be to these early days of mobile what Flying Toasters were to the first color PC monitors -- a simple pleasure that gave us a light entry point to imagining what the new technology could do. But it is a very instructive entry point at that. Crush or Flush reminds us how different the mobile platform is from the Web, where giving people a range of choices and linkage, linkage, linkage to a wide world of options was critical. The disciplines on a handset will involve giving people a very simple tool, binary choices, but letting them drill endlessly. If you narrow the field to a target that is relevant to users, informed by their current desire, then a simple yes/no interface is all that is necessary. The control you give them is not where to go, but how deeply to drill vertically, and when to stop. In the right context, those binary choices are enough even for an age of user-centric media. If it is possible to be both dumb and deep at the same time, then that is the model to which we should aspire.