Friday, May 5, 2006

Foxmarks and the Vision of Places

People who have been following closely the development of the next versions of Firefox -- Firefox 2 and Firefox 3 -- already know that one of the key enhancements of Mozilla's next generation of browsers is a system called Places. Places is based on the key insight that two systems that currently exist separately in the browser, bookmarks (which represent places that you've visited and want to be able to return to) and history (places that you've actually visted), could be much more valuable if they were combined in some fashion. They both exist to solve an end-user problem: how do I get back to some place that I've been before?

As envisioned by its creators, Places integrates history and bookmarks into a single database for the back end, and revamps the current UI affordances (the bookmark menu, the bookmark manager, the toolbar, the bookmark sidebar, the history sidebar, the address bar url history) to talk to that back end. Done right, a system like this really opens up a treasure trove of interesting end-user possibilities, both for extensions and for the core browser. A good interface to the Places back-end would also enable online service providers to provide tight integration -- right in the browser -- with the many different kinds of web services that have exploded in popularity recently, such as, digg, and furl. Despite the fact that any implementation of Places would likely require a major rewrite of Foxmarks, we've been excited about Places since we first heard about it; we have drunk the Kool-Aid, and we believe in the Vision of Places.

That's the good news. The bad news is that, in the rush to get Firefox 2 out the door, the implementation of Places falls quite a bit short of the vision. There have been numerous complaints about the number of bugs in Places in the alpha version of Firefox 2; some have observed that the new UI is not the hands-down improvement that was hoped for; and we noted that the back-end was designed without consideration of synchronization, which would have made a Places version of Foxmarks essentially impossible.

That's the bad news. The good news is that Mozilla has made the tough decision to hold off on deploying Places at least until Firefox 3, giving it some much-needed time to be more fully developed. Some critics have complained about this decision, but we think it's the right call -- something as important as this deserves time for considered development; no one would be served by releasing Places before it was ready for prime time.

Last week, Mike Shaver of the Mozilla Corporation gathered some of the interested parties in a meeting to discuss requirements for Places that hadn't yet been reflected in the implementation. Representatives from Google, Yahoo,, and yours truly discussed both back-end needs and real-world use-cases. I came out of the meeting with my belief in the rightness of the Places Vision reaffirmed, though there's going to be quite a bit of blood, sweat, and tears shed getting there. I look forward to seeing Shaver's notes from that meeting published for all to see.