Is .mobi the Return of the Web Content Ghetto?

In 2003 I built, if not the first, one of the earliest web sites for mobile phones and PDAs at Penn State. Addressing the common criticism of stale (i.e., forgotten, and presumably static) content on separate, mobile-optimized sites at that time, this site was dynamic, pulling its content from the same database used by the main site. The only technical difference was that the templates used WAP2 markup and offered a single column for content, simplified navigation, and limited market branding.

Nonetheless, the problem of content ghettos serving mobile devices has contributed to the relatively slow adoption by early users. But these early attempts at least tried to address the limitations of the small form factor and engage the mobilized users. The larger barrier has been, of course, the complete unsuitability of most sites to rendering in mobile devices' browsers - until recently.

Enter dotMobi, an industry consortium to fix what ails the mobile Web, and to innovate for better user experiences. Great, right? Well, maybe, but .mobi introduces new problems of it's own. .mobi had a surge in popularity when the initiative started. And today, .mobi and kindred WAP sites account for about a quarter of the mobile web traffic. But, the trend is down as devices and browsers are increasingly sophisticated and able to provide the rich application functionality of their larger implementations. Traffic to social web sites accounts for about 40-60% of web traffic worldwide, and you can be sure that WAP sites won't deliver the user experience that drives their popularity. We can adapt a yarn from Wall Street here: don't fight the trend.

Moreover, since .mobi is a top level domain, it has already segmented itself from most web resources in education, government, non-prophet and personal domains, i.e., those who are least likely to own and manage additional TLDs. In fact, the mission of .mobi is primarily to serve small and mid-sized business. Even the 4 letter TLD itself is problematic, as on many devices "mobi" requires 9 keystrokes. Why, those same devices could type out "wap" in just 3.

Sites with a .mobi TLD would have access to some powerful applications. For example, a device database service hosted by .mobi will allow fine customization of web sites for any of the hundreds of devices on the markets. If that sounds a little insane to you, I suggest that maybe you're right. (Although, it does sound like a boon for developers ;-). If making content ghettos sounds evil, how will you reckon with ghettos of user experience. I've had similar concerns about customizing markup to optimize for iPhones' Safari browser. And while evangelizing standards and accessibility I railed against code forking during the Microsoft and Netscape browser wars. Yes, these are concerns and my position is inconsistent, but I'm still trying to justify the hassle because the iPhone user interface is exceptional. Confused? So am I, but I'm watching with fascination as standards and best practices emerge. What do you think?