You do understand that IE still has overwhelming market share, and a great majority of our site visitors are using IE?
Which is why I did not disregard IE entirely, but it is far easier, and will result in far cleaner code to maintain, to code first for browsers that at least get the standards mostly right (there'll always be bugs) than for IE which still gets it so very wrong. Adding workarounds for IE at the end will be a far less tricky task in comparison to coding for IE first and supporting other browsers later.
Even without that my points about graceful degradation/progressive enhancement remain valid. There is also the very high likelihood that if you code to standards and hack for IE later you'll see pages with massively reduced sizes, resulting in a nice bandwidth saving (though admittedly probably not as much as some other sites, as the bulk of your bandwidth is probably downloads). See if you can get the Slashdot guys to give you a rough idea of their bandwidth use before and after they switched from their HTML 3.2 layout to the current layout.
Mess of a site?
Take a look at the page source some time if you know your HTML/CSS/JS
For a real world example if you don't know those languages: open up Wincusto in a clean Opera install. Then use site-specific preferences (right-click-->Edit Site Preferences...-->Network tab-->Identify as: Mask as Firefox) to make Opera pretend it is Firefox. Suddenly the site will work a whole lot better, not because Opera is doing anything different, or changing how it implements standards, but because the site's code is not rejecting Opera out of hand anymore. The trouble with this approach is that anyone using it then appears to Wincusto's servers to be a Firefox users – Opera's visible market-share on this site becomes artificially low as a result.
I posted
that advice a day after the current site launched and included simple advice (i.e. treat Opera the same as Firefox in the JavaScript) to the authors that would be trivial to implement and improve (though not perfect) things quickly. This advice about Opera was included (although with the erroneous implication that Opera's behaviour changes with respect to how it handles the code its sent) in a post
promising Opera support (which I note includes a request for Opera user to be patient, but I think I've done that for long enough) somewhere down the line but I don't recall seeing a single change designed to improve the site's performance in Opera. I then posted
updated advice, which would net support equivalent to masking as Firefox, but without that copy of Opera disappearing from the server stats; I was promised this would be added to the note on Opera compatibility but it never happened so it's likely that virtually every copy of Opera bar mine used to browse this site is actually counted for stats purposes as Firefox.
If it sounds like I have an axe to grind that's probably because I do. I've waited (even offered my time to help) near on two years for
any kind of promised improvement only to see nearly bugger all. Right now I won't subscribe because it seems my money will be wasted on excessive bandwidth and site code 10 years out of date. I really like Stardock as a company otherwise though, I even have an OD ultimate subscription.
