I was blown away when I came back last night a few hours after posting my request for help, to see how many of you had responded.

Now I have to impose on you further and ask you to do it again. The test results didn't individually give me a clue to the problem, but the fact that so many people experienced exactly the same problem was itself a clue. To test my theory about what's wrong, I've created a separate test site: http://test.msmphotography.com/ . This should be pretty simple: It's either fixed, or it's the same thing as before.

If you do experience problems, I'd like you to append "?debugmode=true" to the address and refresh. This will split the browser window in two, and on the right will be a long list of technical gibberish. The important thing is to look for error messages in red, at the end of the list right where things went pear-shaped. There will be a few error messages above that point, but these should be "normal", in that they're benign errors that are detected and worked around with alternatives (yeah, real world code is sometimes like that).

FYI, the theory we're testing is that it's a timing problem. The standard design I've used for a long time loads small packages of JavaScript depending on the browser type; i.e. browser-specific script. The thing I realized last night was that I've always tested it under exactly the same conditions, either here in my office on the office net, or to the public server over broadband...and the timing is therefore always exactly the same. But if these secondary script modules haven't arrived by the time that the page and its main script are ready to rock-and-roll, ka blam!. I first saw this testing Safari, interestingly enough, and connected the dots to wonder if it wasn't happening to other browsers too.

Or maybe this is another theory to be tossed on the ash heap of history. Time and a dozen Zoners will tell.

Thanks in advance for once more going into the breech.