UIUC security researchers apparently feel that no amount of work on existing browser will be sufficient so have set out to design a new one with security built-in from the start. The browser, OP, first and foremost breaks down into a core set of modules and security policies governing the interactions between them. This seems like a good idea in terms of compartmentalizing risk when vulnerabilities eventually crop up. They also seem to agree largely with Firefox’s approach to adding features intending to enhance security for 3rd party plugins as well without demanding more of those developers. I especially like their approach to transparency, where apparently there will be some capability to backtrack security problems and identify from what site they originate.
I wish this research team had more strongly considered working with an existing open source browser, Firefox in particular. They are using KHTML for OP and apparently plan multiplatform support using WebKit but while they may get ahead on security, how much are they sacrificing on other fronts to do so?
Given the recommendations to use a second, different, more secure browser I saw some time back for accessing sensitive sites, I think OP will have a place in my toolkit, regardless, as it seems to be a great fit just for that.