It's been a year, so it's time to revisit. I'll describe how that particular area has evolved, and whether it was a good call or a bad call to focus on that area of expertise.

• Dampening: Also called damping, I described this as "the softening of a user interface through gradual transition instead of immediate state changes". What was considered fairly new then is pretty much standard practice now. I wasn't explicit about how to take advantage of the opportunity, but it seems today that the hard part isn't creating the effects: Components such as the Prototype framework and the Scriptaculous library simplify the development process. The challenge is in identifying how and when to make good use of the technique. Verdict: Good call. Experts in this area are even more valuable than they were a year ago.

• E4X: As I defined it last year, "[s]mart, sensible handling of XML in Javascript" for users of Firefox or Flash. It's a great, powerful technology, but it just hasn't taken off, particularly because its client platforms still don't include native support in Internet Explorer or Safari. Verdict: Bad call. It's still worth hoping it comes alive in the future.

• JSON: JavaScript Object Notation is the cat's pajamas when it comes to exchanging data in a lightweight manner. I had some concerns that people would see JSON as competing with XML as a format, but they turned out to be unfounded. JSON has taken off for many APIs, including a complete developer center from Yahoo, which covers the JSON APIs available for Yahoo services like del.icio.us, Search, Travel, Answerss and nearly every other Yahoo service. It's also shown up in places like the AOL Pictures API and on Eventful. Our team at Six Apart has used it extensively to power Vox, and Tatsuhiko Miyagawa also built it into his Plagger platform. This one's a gimme. Verdict: Good call. JSON skills are a must-have for serious Ajax hackers.

• Good ole' XHTML and CSS: Okay, I cheated on this one. But it was mostly serving as a reminder that you can't forget the basics. Verdict: Always a good call.

• Buffering: The idea here is that the big gobs of Javascript that power Ajax apps would require accommodations for the time they take to load. Techniques like pipelining and intelligent caching have helped mitigate this need somewhat, but there are still valid concerns about the user experience and performance challenges involved in creating rich applications. Verdict: Still out. Can't hurt to be up to speed here, but it's not a deal-breaker if you aren't an expert yet.

• The Atom API: The potential for building on top of robust API for data storage is enormous. We've seen some fits and starts and progress here, but the Atom API hasn't gotten finalized in the way that feed format did, and that has somewhat affected adoption. On the other hand, we've seen some ringing endorsements: Dewitt Clinton, former principal engineer at Amazon's A9, said, "I’ve found the Atom 1.0 standard to meet the needs of nearly every single problem that I’ve thrown at it." Google's Data APIs for Calendar, Blogger and Base are built on top of the API as well. For general data exchange, the Atom API is strong. For regular posting to blogs, the lack of finalization has meant that Atom will probably show up somewhat later in new tools using blogging APIs, such as Microsoft's Word 2007 and Windows Live Writer, or Adobe's Contribute. Verdict: A (moderately) good call.

• Helping Ruby Grow Up: There's been a lot of progress in this area. A year ago, internationalization and localization on the Rails platform could be painful, but updates to both the core platform and to the applications running on top of it have simplified this work. Best of all, there's a smart, defined path for scalability on Rails applications. As DHH himself recommends, just use LiveJournal's open source infrastructure. memcached is free, recently-updated, and works like crazy. So the immediate need was met by people who saw the value of this opportunity last year, but now there's new issues to tackle. Verdict: Good call.

• Marketing: Another gimme. It never goes out of style, and it's still underrated by most geeks. Learn it, live it, love it. Verdict: Always a good call.

The Bottom Line

So overall, how'd we do? Out of eight predictions, we've got two items that are always a good call, which don't really count as particularly prescient. The verdict is still out on one. One was a bad call, and four were good calls. If you throw out the two that don't count, that's four predictions out of six, with one that could still be valuable in the future. That's a 2/3 chance you would have learned something useful by gambling on those recommendations a year ago.

I'm still working on some new recommendations, but I'm more than willing to hear feedback on my analyses above, especially if you disagree. Thanks to Doug van der Molen for the reminder to revisit.