Re: [whatwg] Character-encoding-related threads

2012-06-29 23:42, Ian Hickson wrote:

> I consider all "boilerplate" to be a significant burden. I think there's a
> huge win to making it trivial to create a Web page. Anything we require
> makes it less trivial.

It's a win, but I'm not sure of the "huge". When learning HTML, it's an 
important aspect, and also when typing HTML by hand, but then it's 
mostly a convenience - and it helps to avoid annoying problems caused 
e.g. by making a single typo in a DOCTYPE declaration. So <!DOCTYPE 
html> is really an improvement

> Currently you need a DOCTYPE, a character encoding declaration, a title,
> and some content. I'd love to be in a position where the empty string
> would be a valid document, personally.

Is content really necessary? The validator.nu service accepts the following:

<!DOCTYPE html><title></title>

I don't think we can get rid of DOCTYPE anytime soon, as browser vendors 
are stuck with DOCTYPE sniffing.

But the <title> element isn't really needed, and unless I'm mistaken, 
the current rules allow its omission under some conditions - which 
cannot be tested algorithmically, so conformance checkers should issue a 
warning at most about missing <title>.

It might be better to declare <title> optional but strongly recommend 
its use on web or intranet pages (it might be rather irrelevant in other 
uses of HTML).

Yucca

Received on Friday, 13 July 2012 13:10:03 UTC