absolute URIs for "rel" attributes and the W3C validator

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
According to RFC 5988, should the following be considered valid HTML5?

<!DOCTYPE html>
   <meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
   <link rel="http://example.com/example-rel-1 " href="/example-1">
   <a rel="http://example.com/example-rel-2 " href="/example-2">example2</a>

It fails to validate using the W3C validation service. Have I misunderstood the
requirements of RFC 5988 regarding non-registered "rel" values (i.e. they must
be absolute URIs)? Or does W3C's validator not fully conform to RFC 5988?

Michael Bradley, Jr.

Re: absolute URIs for "rel" attributes and the W3C validator

In comp.infosystems.www.authoring.html,
Quoted text here. Click to load it


RFC 5988 defines a "Link:" header, not a "<link>" tag. The examples in
the document should clue you in that this is not what RFC 5988 is

not going to look up the exact spec of the "<link>" tag at the moment

Re: absolute URIs for "rel" attributes and the W3C validator

Quoted text here. Click to load it

But RFC 5988 says that it's intended to be independent of the specific

More specifically, it mentions the <LINK> element in a few places:

5.  The Link Header Field

   The Link entity-header field provides a means for serialising one or
   more links in HTTP headers.  It is semantically equivalent to the
   <LINK> element in HTML, as well as the atom:link feed-level element
   in Atom [RFC4287].


Appendix A.  Notes on Using the Link Header with the HTML4 Format

   HTML motivated the original syntax of the Link header, and many of
   the design decisions in this document are driven by a desire to stay
   compatible with these uses.

   In HTML4, the link element can be mapped to links as specified here
   by using the "href" attribute for the target URI, and "rel" to convey
   the relation type, as in the Link header.  The context of the link is
   the URI associated with the entire HTML document.

Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***

Re: absolute URIs for "rel" attributes and the W3C validator

On Friday, September 14, 2012 7:13:32 PM UTC-5, Eli the Bearded wrote:
Quoted text here. Click to load it

It turns out the question I should have asked first is whether the HTML5 spec
does or does not conform to RFC 5988, either implicitly or explicitly, with
respect to "<a>" and "<link>" tags.

Having realized that (thanks to folks in #html on Freenode), I discovered that
in the past year there was an editorial decision to modify the HTML5 spec to
implicitly conform. The actual patch to the spec's source took place last month:


So now it's a matter of the W3C's validation service being updated accordingly.
Anyone have any idea how long it usually takes for a change like that to ripple
down into the validator?

Michael Bradley, Jr.

Re: absolute URIs for "rel" attributes and the W3C validator

2012-09-15 4:10, Michael Bradley, Jr. wrote:

Quoted text here. Click to load it

"HTML5 spec" is vague concept.

The latest published version of W3C HTML5 "specification",
http://www.w3.org/TR/html5/ , is dated 29 March 2012 and appears to be
RFC 5988 ignorant and implicitly forbids absolute URLs as rel values,
because it defines a set of values and normatively (!) cites the
"microformats wiki existing-rel-values page", which in turn does not
allow such values.

Quoted text here. Click to load it

In the editor's draft only, and reluctantly, if I can read between the
lines. It does not mention RFC 5988 at all, and absolute URLs as values
pop up rather abruptly. The text says that if a rel value contains the
colon ":", it must be an absolute URL. There's also the somewhat cryptic
note: "Even URL-valued link types are compared ASCII-case-insensitively.
Validators might choose to warn about characters U+0041 (LATIN CAPITAL
LETTER A) through U+005A (LATIN CAPITAL LETTER Z) (inclusive) in the
pre-case-folded form of link types that contain a colon."

And the following statement means that it's _not_ the idea that authors
can use URLs in an open-ended manner:

"Conformance checkers may use the information given on the WHATWG Wiki
MetaExtensions page to establish if a value is allowed or not: values
defined in this specification or marked as "proposed" or "ratified" must
be accepted, whereas values marked as "discontinued" or not listed in
either this specification or on the aforementioned page must be reported
as invalid."

So an absolute URL would be invalid if it is not mentioned on the Wiki
page. And the draft says what checkers must do, instead of saying what
is correct (or valid); an odd way of writing specifications. I would
even say that this messy approach is a parody of standardization, but
here, as so often, parodies cannot be distinguished from reality.

Quoted text here. Click to load it

It's really a copy of the software at http://validator.nu but not
necessarily an up-to-date copy.

Quoted text here. Click to load it

I cannot give you an estimate. People behind http://validator.nu seem to
the a third force in the HTML5 world, in addition to the W3C and the
WHATWG, so they might decide to do what they regard as best. But if they
read W3C documents, then formally the Working Draft disallows absolute
URLs and the editor's draft is self-contradictory, so what would you do
if you wanted to treat the W3C as defining the standards?

Yucca, http://www.cs.tut.fi/~jkorpela /

Site Timeline