Re: [Fwd: Review of HTML 4.0 Specification]

On Thu, 11 Sep 1997, Misha Wolf wrote:

> David Abrahamson wrote to the editors of HTML 4.0:
> > Each <META> element specifies a name/value pair.  If multiple <META>
> > elements are provided with the same name, their combined contents -
> > concatenated as a comma-separated list - is the value that will be
> > associated with that name.
> 
> I strongly disagree:

For NAME elements, I also disagree for the same reasons.

For HTTP-EQUIV elements, I believe that the HTTP/1.1 specification
requires that elements may be concatenated as stated by David.
Certain servers transform HTTP-EQUIV META tags in HTML into actual
HTTP headers.


> >  * The <META> element should not be used where a specific element,
> > such as <TITLE>, would be more appropriate.  Rather than specifying a
> > <META> element with a URL as the value of the CONTENT attribute,
> > use a <LINK> element instead.

> 
> I don't understand the relationship between the two sentences of this 
> paragraph and also don't understand the first sentence.  I strongly 
> disagree with the second sentence.  "Dublin Core Metadata for Simple 
> Resource Description" (draft-kunze-dc-00.txt) explicitly states that 
> the elements DESCRIPTION, IDENTIFIER and RIGHTS may take URLs as their 
> values.


I think DC is supposed to be transport-independant, while HTML META may be
viewed as a transport, so I suppose one could implement the DC RIGHTS
element as a REL modifier in a link
<LINK REL=DC.RIGHTS SCHEMA=xxx LANG=en HREF=http://x.y/rights.html>
or anchor
<A REL=DC.RIGHTS HREF=http://x.y/rights.html>Copyright</A>
which I personally prefer, but it's confusing to mix REL and META tags
in one metadata set, and embedding URLs in META is enshrined in PICS
and current DC practice.

Andrew Daviel
TRIUMF & Vancouver-Webpages

Received on Thursday, 11 September 1997 17:18:06 UTC