This file is UTF-8 encoded. And a meta element declares it to be UTF-8 encoded.

Now, try this experiement: Manually override the encoding. E.g. set it to Windows-1252.

Then reload the page.

Now the question: Did your Web browser “remember” the encoding that you manually set it to be? Or does it “fall back” to the the declared encoding?

If it remember the encoding, then the question is for how long. In Safari, it will remember it “for ever”, until you close the current browsing context/Tab. Another Wekit browser, iCab, allows you to “unremember” the encoding without closing the Tab. Whereas a browser like Opera seems more unwilling to forget. And so on and so forth.

Personally I like the Firefox way the best: It does not remember the encoding at all — as much as I can see.

PS: HTML5 refers to this as info on “the likely encoding”. Which sounds kind of smart ... But, at least as much as I can see, this step in the algorithm is both quite primitive – and quite wishy-washy with regard to what it is that falls into this cathegory.