2007/7/31, Teo Hui Ming:
>
> But my concern is, can we safely link to any document using atom:link
> (regardless of @rel), just as we can with html:a and html:link, thus
> are "normal links" [1].
>
> -OR-
>
> @rel may imply certain meaning on relationship (e.g. ownership)
> between documents that authors/implementers must be careful about. If
> yes, it seems necessary to make it clear.

atom:link/@rel has the exact same implications as html:link/@rel / html:a/@rel.

The only difference with HTML is that in Atom there is *always* a
'rel' value (someone might argue that in HTML, when there is no @rel
on an html:link or html:a, you can make it default to rel="related").

> and atom:[EMAIL PROTECTED], should we treat it as normal / embedded [2] ?
>
> [1] http://www.w3.org/DesignIssues/LinkLaw.html#Normal
> [2] http://www.w3.org/DesignIssues/LinkLaw.html#Embedded

Embedded, as Eric said.

Hence the @src vs. @href

> > [..] However, rel="related" doesn't
> > really communicate the relationship between the reference and the referent,
> > so I'm tempted to propose a new relationship. rel="referent" seems
> > appropriate.
>
> A new precise relation type is fine for new systems. But existing feed
> readers prefer the first alternate atom:link in feed entry. If service
> implementers need to compromise, at least they know it's okay to do
> so. Which is why, I think, the question above is important.
>
>
> Thanks,
> --
> Teo Hui Ming
>
>


-- 
Thomas Broyer

Reply via email to