This is actually correct behavior. In fact, it is one of the few things that IE has consistently gotten right throughout the years that other browsers have not: ' is _not_ valid HTML.

Please read: http://fishbowl.pastiche.org/2003/07/01/the_curse_of_apos

You’re absolutely right! I use this.quirks whenever I’m in the sprite’s scope. The example above is actually in the onload handler for an image, so the scope is an image object. If anyone’s curious, here’s a link to the sprite source.

Or maybe you do check for them in this.quirks rather than LzSprite.prototype.quirks?

Strong points! You’re absolutely right about Browser detection vs. quirk detection - I avoid browser detection wherever possible. Unfortunately, most of these quirks are things that are difficult to detect with quirk detection, like which direction the mousewheel reports it’s moving. Or, they’re so tied to a specific browser (like AlphaImageLoader is to IE) that I found it more expedient to do browser detection. I’m so used to checking for browser quirks elsewhere (document.all for example) that I never bothered to follow my own advice and pull these tests out into quirks with explicit IDs. It never occurred to me to do this until I was writing the post yesterday - funny how writing about things often clarifies the one’s understanding!

And, you’re absolutely right about how quirk names shouldn’t be named after the browser! This is something that occurred to me, but I ran out of time before 4.0 and it’s seemed like a low priority since.

Thanks for the feedback - I really enjoy blogging this stuff because of the quality comments I get from folks like you!