W3C home > Mailing lists > Public > www-html@w3.org > January 2005

RE: Suggestion: 'rel="unrelated"'

From: Mark Birbeck <mark.birbeck@x-port.net>
Date: Fri, 21 Jan 2005 14:12:50 -0000
To: <www-html@w3.org>
Message-ID: <001901c4ffc3$4dabb220$6f01a8c0@W100>

[In response to the various suggestions for different values of @rel.]


In the first instance -- in relation to the current proposal from Google --
why do we need to mark-up the tag at all? The issue is very specifically
that Google does not want to follow the links in blogs, and it also does not
want to give a higher page ranking to any pages referenced from those blogs.

The goals are fine, but they have proposed a solution to this that is a hack
-- it has the appearance of being a new attribute that we can all use in our
mark-up, but what they are actually going to do is automatically add
@rel="nofollow" to all (most) anchors in a blog, regardless of what was
originally put in.

I say this is a hack because they appear to have gone through the following
steps:

1. Faced with the understandable problem "how do we avoid indexing links
   from blogs", they rightly say, "how do we know which links come from
   blogs"?

2. Faced with this secondary problem they *could* have said "why not add
   some metadata to the head of the document to indicate that it is a blog,
   and then make our crawler behave accordingly".

3. Or they could have changed their blogging software so that any comments
   that are posted that contain links have to be approved by the blog
   owner.

4. Or a hundred and one other solutions ...

5. Instead they said "why not get all those who provide blog services to
   add some mark-up to each link so that the crawler doesn't need to know
   the difference between a blog and other documents -- it can just check
   the link properties".

6. And then of course someone said, "great idea, and let's use the 'rel'
   attribute, to do it with since it's not used much, and no-one really
   understands it anyway, and it makes it look like we've produced a
   generic, standards-based solution".

So I would suggest that any discussion about whether it should be this value
for rel or that value are pointless, since that isn't what is on the table.
What we have is Google marking up its own documents (they host many of the
blogs) in such a way that it can get back information about them, and
they've done it with a kludge.

Regards,

Mark

PS Lachlan -- none of the above is to say that your proposal for linking
metadata is not useful or important, it's just that unfortunately the
'nofollow' proposal isn't operating at that level of subtlety.


Mark Birbeck
CEO
x-port.net Ltd.

e: Mark.Birbeck@x-port.net
t: +44 (0) 20 7689 9232
w: http://www.formsPlayer.com/
b: http://internet-apps.blogspot.com/

Download our XForms processor from
http://www.formsPlayer.com/

> -----Original Message-----
> From: www-html-request@w3.org 
> [mailto:www-html-request@w3.org] On Behalf Of Lachlan Hunt
> Sent: 21 January 2005 11:31
> To: Jens Meiert
> Cc: ic@rimantas.com; www-html@w3.org
> Subject: Re: Suggestion: 'rel="unrelated"'
> 
> 
> 
> Jens Meiert wrote:
> > Nonetheless, what about 'rel="unrelated"'?
> 
> If a resource is unrealted, then why link to it in the first place? 
> Unless, of course, it is a result of a user contribution; 
> however, there 
> is little chance that an automated publishing tool handling that 
> contribution (without moderation) can tell the difference between 
> related and unrealted (spam) comments, and thus cannot apply the 
> relationship appropriately anyway.
> 
> Personally, I'm leaning towards rel="unendorsed" at the moment.  It's 
> definition could be something like:
> 
> Unendorsed:
>      Refers to a resource that may be related to the linking
>      document, but is not endorsed by the author. This should
>      not be counted as a negative rating for the resource,
>      however.
> 
> User agents that choose to issue credit to a resource based 
> on the links 
> to it should issue a lowered, or no, credit rating for the link. 
> However, unendorsed (even though it has a similar effect to 
> nofollow) is 
> not intended to be applied to all links recieved through user 
> contribution; instead, I'm thinking that rel="contribution" and a few 
> others are more appropriate for those use cases.
> 
> I am currently working on a draft proposal called "Web Communication 
> Link Relationships" (WCLR) that I'm considering sending off 
> to the GMPG 
> [1] to produce an XMDP, or to the W3C if they're interested.  
> It should 
> be finished and published in a day or two and the aim is to define 
> relationships that will facilitate web communication through 
> increased 
> linking semantics.
> 
> It's currently split into 4 categories including:
>    1. User Contribution
>    2. Resource Tracking
>    3. Communication Tracking, and
>    4. Endorsement
> 
> I realise some of those category names may not make a lot of sense to 
> you now, but they do when explained in the draft that I'm currently 
> writing.  The relationships are based on the various ideas several 
> people have published in their own blogs over the past few 
> months and a 
> few of those that I published earlier [2] (just some from the "User 
> Feedback" and "Endorsemnt" categories in that post).
> 
> Rimantas Liubertas wrote:
> >  Why not just rel="nopagerank"?
> 
> Because a relationship should not express user-agent-implementation 
> specific functionality, but instead a semantic relationship between 
> resources.
> 
> [1] http://gmpg.org/
> [2] http://lachy.id.au/blogs/log/2004/08/link-relationships
> 
> -- 
> Lachlan Hunt
> http://lachy.id.au/
> http://GetFirefox.com/    Rediscover the Web
> http://SpreadFirefox.com/   Igniting the Web
> 
> 
> 
Received on Friday, 21 January 2005 14:13:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:01 GMT