[microformats-discuss] rel="cite", rev="cite". rel="comment"?

[Andy Skelton]

On 7/13/05, Tantek Çelik <tantek at cs.stanford.edu> wrote:
> On 7/12/05 12:12 PM, "Andy Skelton" <skeltoac at gmail.com> wrote:
> What semantic difference is there between:
> 
> 1. <cite><a href="http://example.com">example citation</a></cite>
> 
> and
> 
> 2. <a rel="cite" href="http://example.com">example citation</a>
> 
> ?
> 
> AFAICT, there is no semantic difference.

I agree.

> However, (1) only uses pre-defined XHTML elements.
> 
> (2) uses a new rel value.
> 
> Thus (1) is preferred.

Also agreed. Furthermore, looking forward to times when hypertext may
be more than just <a href></a>, people should get accustomed to
semantic use of all valid elements.

> > and rev="cite" to hyperlinks.
> 
> I don't think rev="cite" is something that an author can be trusted to
> claim.
> 
> From a trust and authority perspective, you don't get to state that someone
> else is citing you.  You can only state when you cite someone else.

I see the trust/authority issue in that. Still, there should be a way
to indicate the reciprocal relationship from a cited resource. It's
something to consider.

> While I can certainly appreciate the symmetry of using rel vs. rev = cite,
> in practice, 'rev' is something that is not easy to understand for web
> authors.

The 'rev' attribute could be more clearly documented but that is not
an excuse to treat it like it's deprecated. The power of 'rev' is
begging to be exploited, not feared.

<link rel="Next" href="page2.html"> means "yonder document is the next
document in the sequence."

In cross-domain linking, <a rev="whatiam" href></a> gives the author
of the referring document the power to define his document's role in
the relationship. "The current document is a comment on the linked
document." "This is a rebuttal to that." "This is evidence to support
that."

> Another alternative is to use something like rel="comment", which would be
> like saying that this resource over here, is a comment for the current
> (portion of the) page.

That works on the other end of the same vector--the pingback receiver.
The off-site comment author would use rev="comment" to describe his
own document. Maybe we don't need a "cite" value for rel/rev because
we have the cite element but we could make good use of some other
rel/rev values to map discussions across domains.

Andy