2006/11/30, Mark Baker:
The real problem here AIUI - at least in the context of HTML 5's inferred rel="feed" bit - is not just entry documents, it's any Atom document which wouldn't normally be considered a "feed" by a typical user; something that most people would be interested in subscribing to. An example I gave on the whatwg list was an MHTML-like (MIME multipart) package, but there are many other possible examples of course; not all RFC 4287 feed documents are "feeds" in this sense.
Yes.
If HTML 5 (and current practice) doesn't change, but we defer to them for the specification of autodiscovery, then a new media type would be one way forward. But it should be reusable for all non-"feed" (i.e. from a user POV, as above) Atom documents, not just entry documents; perhaps application/atom-no-feed+xml. It's an ugly hack, but it's better than the alternative of many more specific Atom-related media types, which atomentry+xml might set a precedent for.
-1 This means RSS would need two media types. This also means an HTML document can be a "feed", or it needs its own media type. In an "entry" page in a blog-like scenario, we could find: <link rel="feed" href="/feed.atom" type="application/atom+xml;type=feed"> <link rel="feed" href="/" type="text/html"> which tells you the page "believes" it is an "item" in the linked "feeds". Then , in /feed.atom: <atom:link rel="alternate" href="/" type="text/html" /> and in the "/": <link rel="alternate" href="/feed.atom" type="application/atom+xml;type=feed"> And yes, HTML pages are "subscribable", either using a microformat (see the W3C's home page, from which the RSS is actually produced using an XSLT stylesheet), or using user-defined "scrapping" (some aggregators allow you to subscribe to any web page; they try to infer the "entries" from the semantical markup –h1, h2, h3, etc.– and you can customize the mechanism in per-feed basis: this page uses h2, this one uses h3 with class="article", etc.) The relation from the "entry" page to the feed or "home page" is the same (hey, they are alternates!): they are "feeds", whatever their representation (Atom, RSS, HTML, etc.) The difference with "container"? a "feed" is known to have a representation which only exposes a subset of the contained items. This is the same as: <link rel="contents" href="/toc.html" type="text/html"> <link rel="contents" href="/toc.opml" type="what's the media type for OPML?">
Another way forward, because the rel="feed" inference is triggered not just by the media type but by the "alternate" relationship keyword, is to create a non-feed alternate relationship ("alternate-non-feed"? ick). I prefer the new relationship to a new media type because it's less disruptive; it doesn't require futzing around with existing specs and implementations.
I'd prefer basing autodiscovery on the media types and not at all on the relationships. A "feed" relationship would only help finding the "living resource" (similar to rel="current" in the Atom Relationship Registry) if you're not already "on" it (in which case, rel="alternate" would be used). UAs would then obviously continue to support autodiscovery using "alternate" all-over-the-place, this would just be a lucky side-effect; and everyone would be happy. -- Thomas Broyer