Currently I am using $.browser.webkit and $.browser.msie because the input types between the browsers have plenty of unique UI differences that I am having to accommodate for.
Specifically: Chrome changes my range to a slider, IE doesn't; and I have to place a spinner on the IE time type, Chrome covers that. There are others, but these are the ones that I have to deal with the most.
What is the preferred way to test for these differences rather than using the migrate lib's browser test? (I know I could query the useragent directly, but I would rather query the UI to see if I need to add a spinner, or slider, etc.)
Short answer, mplungjan had an answer that was close but not what I needed. If you look at IE11 vs Chrome where both have a feature that is supported, you may see where they render the support completely differently. So my solution ended up being stay with the migrate lib to detect browser and version, then have my code override the render on browsers that need it.
This means that I will be keeping a dictionary of browsers and implemented features, so my code can intelligently override as dictionary dictates.
As for the case where a feature isn't supported, I will just continue to use the standard test for feature, that has worked and mplungjan has linked to one of the easiest ways to test. Not the best, but the easiest.
Switching to Twitter Bootstrap and Angular.JS UI has reduced some of this oddness, by using other peoples solutions.