Here's a post from John's blog in which he touches the topic of a general strategy for browser support: http://ejohn.org/blog/the-browsers-of-2009/
He also briefly writes about it in his (latest?) book: http://www.manning.com/resig/ Finally, John's (and thus jQuery's) testing strategy is to test against the previous, the current, and the upcoming (nightly build) version of supported browsers. [citation needed] :p P.S.: When I write 'John', it's probably really the whole team behind jQuery that worked out those approaches. On Jan 12, 12:41 pm, RobG <rg...@iinet.net.au> wrote: > On Jan 12, 1:24 pm, Nathan Klatt <n8kl...@gmail.com> wrote: > > > > IE 6 use is 3 times that of Safari (all versions) depending on whose > > > statistics you believe. Why not drop support for Safari while you're > > > at it? And Opera and Chrome? > > > Because you don't have to do anything to support Safari or Chrome or > > Opera > > There are at least 5 Safari-specific quirks catered for in jQuery (I > just searched for Safari in comments), there are likely others. > Admittedly that's far fewer than are required for IE, but since IE 6 > is now about 10 years old, surely it's quirks are well known and > catered for? > > > - they actually work. To stop supporting them you'd have to stop > > supporting standards. > > Browsers will continue to evolve. If appropriate feature detection is > already in place and effective alternatives provided to handle quirks, > you may find that you are handling some new quirks without having to > write a single line of code. :-) > > For example, while the bit in jQuery that checks if the event.target > is a textNode is meant for Safari (pre version 3 I think), it will > work for any browser that has such behaviour. Safari's behaviour was > actually compliant with the spec, the far more common behaviour (i.e. > that event.target is always a nodeType 1) is not compliant. > > > >> I work with several clients that do > > >> not want to "lead the way" in this respect, and need to support IE6 as > > >> long as it has a fair usage share, which may be for several more > > >> years.] > > > > That is a sensible decision > > > Anyone clinging to IE6, at this point, has gone waaaay beyond not > > leading the way! > > If w3schools' statistics are at all accurate, there are about the same > number of people using IE 6 as either IE 7 or 8. > > Of course "support" might mean whether new functionality is provided > for old browsers and whether they continue to be part of a test suite. > New functions that aren't tested in old browsers can simply be marked > in the documentation, or simply "features add after version x.y have > not been tested in browser X" so users know not to use them. > > -- > Rob