On 1/18/06, Charles Iliya Krempeaux <[EMAIL PROTECTED]> wrote: > > > No, but <a rel="alternate" type="application/xml">foo</a> is. > > > > > > RTFS: http://www.w3.org/TR/REC-html40/types.html#type-links. > > > > Excellent - thanks Mark and Ryan! I didn't realise that the "type" > > attribute was valid on <a> elements. > > > > Does that require that we serve the XML with the application/xml MIME > > type? The spec calls the type attribute an "advisory hint"; I'm not > > sure whether this means it has to be correct, or only that it has to be > > close enough (e.g. to text/xml or application/something+xml) to let a > > reader know that the linked resource is XML.
If the "XML source" has a registered MIME type (like, say, "application/atom+xml", as defined in RFC 4287): * Ideally, you should serve the XML source page with the full registered MIME type. * If you want to do strange things with the XML source (like being able to view it in a browser, possibly styled with CSS and/or transformed with XSL), you are allowed to serve the XML source page with the general "application/xml" MIME type. * Either way, you should use the full registered MIME type in the @type attribute when you link to the XML source. <a rel="alternate" type="application/atom+xml" href="..."> (substitute the MIME type you're actually using). If the "XML source" is just some random XML document in a format that you made up, or some format that doesn't have a registered MIME type (like RSS): * Serve the XML source page as "application/xml". Do *NOT* just make up a MIME type, unless you know how to do it properly. * Link to it with <a rel="alternate" type="application/xml" href="...">. > Wasn't "text/xml" deprecated?! And instead you are suppose to use > "application/xml" or "application/???+xml", etc. (If I remember > correctly, this was done because of text-transcoders and conflicts > between "text types" specified with HTTP and XML. I think I read this > on Mark Pilgrim's blog... but I couldn't find the article.) Don't use "text/xml". Ever. Here's why: http://feedparser.org/docs/character-encoding.html -- Cheers, -Mark _______________________________________________ microformats-discuss mailing list [email protected] http://microformats.org/mailman/listinfo/microformats-discuss
