Re: siteData-36: strawman

Tim bray has written:
>You could easily include this in the <head> of HTML documents along the
>lines of
>
>   <meta http-equiv="Web-site" content="http://example.com/site" />
>
>Perhaps <link> would be better, or perhaps the HTML people might want to
>define new markup for the purpose.

I would encourage to use a mechanism which already exists more than 
to define a new one. It will help people to implement it easily in 
HTML 3.2, HTML 4.01, etc.
So the meta and the link seems to be the good solution.

Now, there's something which is a bit tricky. What's a Web site and 
how to define it?

Let's identify some of the possibilities:
1. Domain name = Web site
	http://stars.example.org/
	This Web site is completely dedicated to stars and is 
maintained as whole unity... (whatever it means?)

2. Host = Collection of Web sites.
	http://shelter.example.com/
	This Shelter Inc. company has many Web sites.

	http://shelter.example.com/lovebug/
		The Web site of miss Lovebug, diary, weblogs, etc.

	http://shelter.example.com/cow/
		The Web site of someone passionated by cows.

3. Distributed Web site.
	Imagine now that the notion of Web sites is a collection of 
resource on a topic that people has agreed to put together.
	Example: an encyclopedia on the Web where you have 
distributed specialists, each of them on his specific domain name or 
even individual Web page.

	They have all agreed to participate to the same project and 
they want that an autodiscovery mechanism, directory, etc help them 
to put together the resources.

	There's a need for identification to a home page or a 
resource on the Web, the proposal for a link or meta is nice but has 
a big issue.

	Abuse. Exactly like the keywords system in meta.
	People will start to declare that they belong to a Web site 
to be automatically indexed in the system which put the resources 
together.

	So if you have an URI which says we are all part of this Web 
site, Web site in this case meaning the 3rd option. What will be at 
the end of this URI ? A mechanism which says: "Hey yes, all this list 
of URIs belong to the same site."  ?



-- 
Karl Dubost / W3C - Conformance Manager
           http://www.w3.org/QA/

      --- Be Strict To Be Cool! ---

Received on Friday, 28 February 2003 18:39:51 UTC