html header and html body

Printable View

11-10-2008
Checker1977

html header and html body

If a page does not have header and body element, in the RFC standard, we should treat the information as header or as body?

- I tested with IE and IE treats as body;
- I read from Http 1.1 RFC, but found not covering this topic.

Any insights?

I give a sample here.

Code:

<html> Hello, the information here should be belong to header or body? </html>
11-10-2008
Akkernight

... Why even use that o.O? It ain't hard to type <body></body> or <head></head> :P
11-10-2008
Shakti

Because if somebody is writing a browser (for fun or otherwise) its good to handle cases like those.
11-10-2008
Salem

HTML has no sense of being "well formed", so browsers have to deal with all sorts of crappy broken pages in a variety of ways.

Try stuff in Opera / Firefox / IE and take the majority.
11-10-2008
CornedBee

The HTML5 spec-in-progress has detailed error recovery specifications that are based on current browser behavior. It's probably your best source of information.

http://www.whatwg.org/specs/web-apps...ork/multipage/
Specifically:
http://www.whatwg.org/specs/web-apps...g.html#parsing
11-11-2008
Checker1977

It is Tech Board. We can discuss anything weird.

Quote:

Originally Posted by Akkernight

... Why even use that o.O? It ain't hard to type <body></body> or <head></head> :P
11-11-2008
Checker1977

CB and others, I have tried in browser, I have 3 browsers, IE8, Firefox 3 and the Google Chrome. I also have powershell at hand if you call it is a browser which could retrive web content.

My question does not matter about how browser treats it, I just want to know what is the Http standard defines for the things if not in header and body explicitly, what should they be, belongs to header? body? or anything else?

You can think I am weird, and I like collecting weird pages and post to blog.

Quote:

Originally Posted by CornedBee

The HTML5 spec-in-progress has detailed error recovery specifications that are based on current browser behavior. It's probably your best source of information.

http://www.whatwg.org/specs/web-apps...ork/multipage/
Specifically:
http://www.whatwg.org/specs/web-apps...g.html#parsing
11-11-2008
CornedBee

The HTTP standard doesn't define anything, of course.

The HTML standard says that if you don't have explicit html, head or body elements, the rules are:
1) A html element is automatically wrapped around everything.
2) An empty head element is automatically created.
3) Everything not contained in an explicit body or head is wrapped in a body element.
11-11-2008
Checker1977

CB, the only thing I read is RFC2616, Http standard. What do you mean HTML standard??

Quote:

Originally Posted by CornedBee

The HTTP standard doesn't define anything, of course.

The HTML standard says that if you don't have explicit html, head or body elements, the rules are:
1) A html element is automatically wrapped around everything.
2) An empty head element is automatically created.
3) Everything not contained in an explicit body or head is wrapped in a body element.
11-11-2008
Checker1977

I did some search, do you mean here?

http://www.w3.org/TR/REC-html40/

But where is your 3 conclusions comes from? I read related Http header/body part, but no such 3 rules below.

http://www.w3.org/TR/REC-html40/struct/global.html

Quote:

Originally Posted by CornedBee

The HTTP standard doesn't define anything, of course.

The HTML standard says that if you don't have explicit html, head or body elements, the rules are:
1) A html element is automatically wrapped around everything.
2) An empty head element is automatically created.
3) Everything not contained in an explicit body or head is wrapped in a body element.
11-11-2008
CornedBee

The rules are subtle. If you look at the specification of the HTML, HEAD and BODY elements, you'll find that they specify both start and end tags as optional. This means that the tags can be omitted completely. It's up to the parser to determine how the additional content is nested.
Luckily, the only elements that may appear in both head and body are OBJECT and SCRIPT, so it's generally pretty easy to sort out.

(So my 2nd and 3rd rule are inaccurate. TITLE, META, STYLE, LINK and BASE are put into HEAD automatically.)

But this is all theoretical. No actual browser implements SGML. Use HTML5's rules instead.
11-12-2008
Checker1977

Thanks CB. I did read Html 5 Spec today.

http://www.w3.org/html/wg/html5/#the-header-element

I can not find out where you mentioned both header and body tags are optional. Could you let me know which section do you mean please?

Quote:

Originally Posted by CornedBee

The rules are subtle. If you look at the specification of the HTML, HEAD and BODY elements, you'll find that they specify both start and end tags as optional. This means that the tags can be omitted completely. It's up to the parser to determine how the additional content is nested.
Luckily, the only elements that may appear in both head and body are OBJECT and SCRIPT, so it's generally pretty easy to sort out.

(So my 2nd and 3rd rule are inaccurate. TITLE, META, STYLE, LINK and BASE are put into HEAD automatically.)

But this is all theoretical. No actual browser implements SGML. Use HTML5's rules instead.
11-12-2008
CornedBee

The very section you mentioned above:
http://www.w3.org/TR/REC-html40/struct/global.html
11-16-2008
Checker1977

Hi CB, do you mean section "4.4.7 The header element"?

http://www.w3.org/html/wg/html5/#the-header-element

I read it a couple of times but can not find the logics you mentioned,

1. header and body are both optional tags of Html;
2. if content is neither in header nor in body, it is treated as in body?

Quote:

Originally Posted by CornedBee

The very section you mentioned above:
http://www.w3.org/TR/REC-html40/struct/global.html
11-17-2008
jwenting

Quote:

Originally Posted by Checker1977

CB, the only thing I read is RFC2616, Http standard. What do you mean HTML standard??

You do realise that http and html are two completely different things?
The http RFCs don't deal with html, nor do the html specifications deal with network transport protocols.
11-18-2008
Checker1977

But how does standard say whether content not wrapped in header and also not wrapped in body belongs to header or body?

Quote:

Originally Posted by jwenting

You do realise that http and html are two completely different things?
The http RFCs don't deal with html, nor do the html specifications deal with network transport protocols.
11-18-2008
matsp

Quote:

Originally Posted by Checker1977

But how does standard say whether content not wrapped in header and also not wrapped in body belongs to header or body?

I thought cornedbee already answered that: Unless it is SGML, it will be treated like it had a body around it.

--
Mats
11-23-2008
Checker1977

The problem is I did not find from which paragraph of the Html Spec 5, the conclusion "it will be treated like it had a body around it" is reached. Any ideas?

Quote:

Originally Posted by matsp

I thought cornedbee already answered that: Unless it is SGML, it will be treated like it had a body around it.

--
Mats
11-23-2008
CornedBee

I quoted the HTML 4 spec. It says that for all of html, head and body, both the start and the end tags are optional. So the elements exist whether the tags are there or not, and all that's left is to correctly sort the subelements into head and body.

For the HTML 5 spec, look at 8.1.2.4 for when it's allowable to omit the start tags of html, head and body.
For parsing, look at 8.2.3.1 and 8.2.5.5 and following. Note that the note in 8.2.5.6 is incorrect.