There are a few interesting ideas (most of them are not new, but it never hurts reminding them). As Joseph Smarr explains, when developing a new version of Plaxo, they found many performance issues that put in jeopardy the release of their new system.

It looks like at first they tried to develop their application without paying too much attention on browser’s limitations, and without considering javascript’s strengths and weaknesses (too much OOP boilerplate, as he says, like getters and setters and stuff…). However they were able to adopt new and better practices, that allowed them to ship Plaxo 3.0.