[uf-discuss] OpenSearch

[Edward Summers]

On Aug 30, 2006, at 12:28 AM, David Janes wrote:
> This looks very possible. In particular,
>
> (1) according the article, Yahoo only returns its results in HTML, so
> we know HTML is good
> (2) one could add an extra field to the OpenSearch XML (a description
> file about how your results are returned) indicating that the file is
> hAtom
> (3) parties not understanding hAtom can just display HTML
> (4) parties wating to understand hAtom can do it themselves or use  
> a webservice

I'm glad someone posted about this--hAtom and OpenSearch seem like a  
natural combination. It's also very cool to hear about opensearch  
support at yahoo.

I don't think (2) is necessary since the type attribute on the Url  
tag [1] does just this. The OpenSearch spec defers to the MIME type  
registry for valid values of 'type' and 'application/atom+html' isn't  
there. There isn't a MIME type for hAtom is there?

To simplify things couldn't hSearch recommend that 'text/html' be  
used and let clients either choose to use the hAtom or just display  
as normal?  Is there any real benefit to having an opensearch client  
know they can get hAtom in a response? Isn't the whole point of  
microformats that they are discovered in existing human readable web  
data and not explicitly requested?

I don't see any immediate benefit in creating a microformat for the  
OpenSearch Description xml since autodiscovery makes it easy to  
locate and use. But I could be convinced otherwise.

It would be great to have an hSearch page that outlined the optimal  
way to use hAtom with OpenSearch. Is this something that belongs on  
the wiki?

//Ed