I was wondering if there are known or undocumented (i.e. not shared)quirks of element locator strategies, cross browser or not. This is with respect to Selenium 1 or RC.

I noticed a few quirks in my automation efforts so far:

* IE works better with IDs than say "name=value".
* CSS locator might detect or not fail in Selenium command in Firefox but the command action does nothing (click doesn't actually click) whereas exact XPath equivalent works.
* CSS locator tested successfully with Firebug + FirePath in Sizzle locator mode (matched desired elements), but fails to be detected in Selenium.

It seems although slower (in IE), XPath has more stability in locators than CSS.