I've been showing a website to a customer who insists on using IE and found out, to my surprise, that IE7 does not autodiscover the Atom newsfeed.
The feed is linked to within the HEAD element of a valid HTML 4.01 Strict page with <link rel="alternate" type="application/atom+xml" href="atom.xml" title="Atom 1.0">, the link is pointing to the correct URL, the linked feed is a valid Atom 1.0 XML file served as application/atom+xml and contains the correct <link rel="self">.
Firefox, Opera, Safari and IE8 beta 2 all do correctly find out the feed and lighten up the relevant address bar button - it's just IE7 that stays greyed out. Same IE7 had no trouble in discovering newsfeeds elsewhere and to pass the HTML part of DiveIntoMark's autodiscovery test suite with flying colors.
Changing the HREF to a fully qualified URL made no difference, nor did changing the TYPE to a RSS 2.0 MIME or removing the TITLE attribute.
The website address is http://www.monteanalogo.net/.
Any hint about what's wrong here?
I copied your source to my local machine, and IE7's auto-discovery kicked in fine once I changed the href to a full url, rather than a relative one:
<link rel="alternate" type="application/atom+xml"
href="http://www.monteanalogo.net/atom.xml" title="Atom 1.0">
Needing a full URL seems to be an IE glitch, as the RFC on atom auto-discovery states that:
The value MAY be a relative URI, and if so, clients MUST resolve it to a full URI ... using the document's base URI
Microsoft's own publisher's guide also lists the full url in the href tag, but makes no mention of it being compulsory:
Here is an example of Atom Autodiscovery:
<head> <link rel="alternate" type="application/atom+xml" title="your feed title here" href= "http://www.company.com/feedurl.xml"> </head>