Oh and while I'm thinking about it, this is a cool tool but I'm not sure of how much it will do to pressure browser manufacturers into improving. Obviously Steve's suggestions have gone into Chrome but beyond that I wonder how much influence he has...
It already had influence. When the tool came out it helped us to find a regression in redirect caching that occurred from Firefox 2 to Firefox 3. We fixed it and it's back to working again in Firefox 3.1. Steve has good clout - he knows his stuff and is happy to provide easy test cases.
Did anyone else spot Safari 4.0 on Steve's page? Is that some Apple folks trying out their latest offering, perhaps? I'm pretty sure WebKit nightly comes out as WebKit.
Steve is simply labeling the WebKit nightlies as "Safari 4" - it's a little bit heavy-handed but it will, technically, be correct. Probably moving un-released browsers off into another area (Minefield/Firefox 3.1, Safari 4, and IE 8) would make things more consistent.
I suspect the profiler's UA detection code is confused by the way WebKit nightlies just run your installed version of Safari but linked to the newer WebKit framework. (Look inside the WebKit.app bundle: the binary itself is under 100Kb.)
Unless you have installed Safari 4 Developer Preview (available to any body with an ADC account), your WebKit.app User-Agent string will say Safari 3.x, with a nightly revision number. The exact same revision on a system with Safari 4 will have a different string, though. So it's entirely possible to have a newer version of WebKit running in an older version of Safari, and vice versa.
I think any nightlies since the DP was released in August _ought_ to be counted as Safari 4, but some may not be for this reason.
