Web-Native UX

I'd like to address something many in the User Experience community would rather avoid, since many times it may interfere with deploying the latest cool widget or Ajax technique that comes down the pike. I want to talk about User Experience Consistency and the web. Because while standards bodies have come into being to coordinate the development of the cornerstone technologies from which we build interfaces, and any web developer worth her salt pays attention to valid, semantic markup and the very latest in CSS techniques and the newest developments in unobtrusive scripting and REST and microformats, the pace of development in web-wide usability standards has been glacial at best.

I bring this up because I've been noticing a slower adoption rate of highly-usable, widget-heavy, responsive, dynamic, configurable, powerful web applications. My source? Purely anecdotal and completely unscientific, among my friends and family and even coworkers at Zillow, who express frustration and antipathy toward websites for even minor perceived flaws, while clunky interfaces in other, more primitive sites are tolerated and even preferred to their more elegant "web-application-y" counterparts. With the exception of certain Google and Yahoo! applications, many powerful, innovative web apps are being ignored. In the rush to push the browser to its limits, it's easy to lose sight of the end goal: making routine tasks easier for end users, in the most straightforward way possible.

Web developers are a sensitive bunch - the profession long disregarded in the eyes of "serious" programmers. Ajax was to change all of that. And with impressive things now being done in one of the most challenging software development environments, the front-end of web development has finally been able to attract some formidable talent away from server-side, OS, and game development. For the better, I should think, the Web has been gaining ground, not only as a place to exchange information, but as a valid, full-fledged platform for software development. That's the idea behind all those standards out there: eventually, if we clap our hands and work hard enough, the Web might supplant desktop-native applications for all but a few specialized purposes. Soon, your computer will connect into the World Wide Continuum; your data will mingle freely with the data of billions around the world on an indistinguishable platform of desktop/web-hybrid applications, the social utopia of the Web will supplant thick-client, rugged-individualist desktop computing, the Singularity will occur, and we will all live in happy harmony with the universe.

Fact of the matter is, not nearly enough consistency and code reuse is happening on the web. To an extent, that's good. I'd like to see the web remain a wild place that functions as laboratory as much as controlled platform. But too often, problems are approached like they've never been addressed before.

Sure, there are attempts at usability standards out there. But web usability is complicated, and in spite of the best attempts of several javascript libraries, nobody, not even Google, is as slick and consistent as OS-native applications.

When I'm designing an interface, I try to take into consideration three primary concerns:

Familiarity.

If I haven't seen it before, I don't know what it does, and I don't want to use it, and may not even recognize it as part of the UI. This becomes a huge obstacle for innovative interface development - more later.

Consistency.

If a widget looks more than 70% similar to something else, I will expect them to behave the same. This has ramifications beyond your website (duh).

Ease of use.

This is a big umbrella, encompassing everything from accesibility to ergonomics to "enjoyment": does my slider have a big enough click target? Can I elect to use the keyboard to control this thingie, or am I stuck with the mouse? Do I find myself repeating the same action for common tasks? Is my path into common tasks streamlined and foolproof?

Back in the dark ages, before DHTML graduated from the shadows of image rollovers, web interfaces were largely built out of browser-native form elements and links to more pages. Usability was a minimal concern, because layouts were simpler and interaction models were much less ambitious. Form interfaces were easy to manipulate, since they were largely designed to use OS-native widgets and behave comparable to their desktop-app counterparts. Links were all blue-and-underlined, and they all took you to a new page. Now that we've graduated from a website- to a webapplication-based web, however, many users haven't followed along. Widgets that don't look like text-input boxes can be hard to spot; a recent usability study at Zillow found as much. Part of the problem may be that many users simply haven't been exposed to web applications to expect that anything other than straightforward input controls to respond to input events. I'd like to think that's part of our responsibility as web developers: challenge our users to explore, experiment, discover. But it's also our responsibility to keep the guesswork out of our interfaces.

Web usability is a moving target. I don't have answers right now to many of these questions, but I'll be discussing them as they come up. This post was to survey the territory; I hope to be able to explore aspects of this issue in greater detail soon. I'll also be writing about the technical details of implementing a large-scale interface architecture that balances web standards with friendly, usable design. I believe in powerful, flexible user interfaces, but only inasmuch as they empower the user. Gratuitous lightboxes are not welcome!