[whatwg] Inferring rel="feed" from the media type

Ian Hickson ian at hixie.ch
Thu Nov 30 14:31:59 PST 2006

On Thu, 30 Nov 2006, Mark Baker wrote:
> >
> > How about introducing a new rel keyword: nonfeed?
> I just suggested that on the Atom lists.
> > Then one could say rel="alternate nonfeed" and it would be clear that 
> > referenced URL contains an alternative representation that is not a 
> > feed.
> Well, that wouldn't work because agents would still pick out "alternate" 
> and the media type and infer "subscribability" from that. "alternate" 
> would have to be left off.  Hence my (somewhat) serious suggestion of 
> "alternate-non-feed".
> But yah, I think that's a decent way forward.  I'm still not convinced 
> that the harmful practice of inferring rel="feed" is beyond the point of 
> no return, but nobody's chiming in to agree with me, and I don't have 
> the time to do the necessary implementer-herding myself, so oh well.

I agree with Simon on this:

On Thu, 30 Nov 2006, Simon Pieters wrote:
> I don't see why this:
>   <link rel=alternate href=foo.atom>
> ...isn't good enough. It is a hyperlink to an alternate representation 
> of the current document in application/atom+xml format that is not a 
> syndication feed. IMHO introducing new rel keywords is an uglier hack 
> than just omitting the type="" attribute.

It's not clear to me that there is an actual real-life problem here. It 
seems there is a perceived theoretical problem, but we also have a 
theoretical solution which works in practice. Unless the problem becomes 
actually widespread, this seems like it would be solution enough.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

More information about the whatwg mailing list