Friday, November 10, 2006

thoughts on Daring Fireball's review of Stikkit

John Grubers review of stikkit is worth a read, for a number of reasons. First of all, it's thorough on designing an interface, which sounds familiar to us information professionals, right? Or at least, it should. Anyway, Daring Fireball has established itself as a place to watch for excellent interface critiques, with an eye for those details that make all the difference for the overall experience. For instance, the section AJAX vs. PERMALINKS nails the danger of over-using ajax-techniques to repeat the mistake of frame-based navigation in the web1.0 days. Gruber does this by describing exactly what's wrong with non-changing location-bars.

Apart from that, this review is interesting because stikkit is. They're more than yet another backpack or delicious clone. They're trying to do something new: mixing different types of information, and letting the system work out what you're trying to say. This works with simple conventions in the text, rather than knobs and buttons. For instance, type 'at' and a time, and the note is automagically put in the calendar. If it works, it means that the Google Generation users are still willing to learn a vocabulary - against the holy grail of simplicity.

This is a big if. Quoting from Gruber's concluding thoughts:
I remain unconvinced that it’s a good idea in the first place. Stikkit strikes me as a very good implementation of a flawed premise. The main problem is that with an utter lack of UI-enforced structure, it’s hard to get a sense of what the rules are.
When you present features as being magical — “just type a date and it knows it’s an event” — it’s confusing and irritating when the magic runs out.
Translating this to the challenge of re-thinking the academic information flow: if this magic works in stikkit, it will be a must for the Scholary Workbench. I know several researchers who are swear by Backpack. It's a great collaboration tool. I wouldn't be surprised if they move over.

The magic needs to work reliably; it needs to be discovered gradually and with ease; in short, it needs to Just Work®. Time to spend some r&d on this.