| |||||||||||||||||||||||||||||||||||||||||||||||||
Tim Fanelli wrote a helpful piece for syndication newbies called How To: Create an Atom 1.0 Feed. His piece contains a couple small errors and a few potentially misleading bits, though, so I figured I'd mention it.
(I would have posted this as a comment on Tim's blog, but he doesn't seem to have comments enabled.)
- NOTE: You don't need to use explicit namespace prefixes in an Atom feed. It's an option, of course, but not a requirement.
- NOTE: atom:updated isn't necessarily the date/time of the last update... it's actually the date/time of the last significant update, with "significant" being defined by the publisher. Thus, some publishers will change atom:updated every time a typo is fixed, some will use it to note revisions to a "living document", and others will never modify the date at all.
- CORRECTION: Several of us argued for it to be mandatory, but in the finalized Atom 1.0 spec, the feed-level "self" link is optional. You are strongly urged to include it, but Tim suggests that you're required to provide it, and you're not.
- WARNING: Using a permalink as an atom:id can be tricky business... think very carefully about your implementation before you do it. If you decide to move your blog from foo.com/blog to blog.foo.com, what will that do to your atom:ids? If you change your permalinks from using query-strings to using "search engine safe" URIs, what will that do to your atom:ids? Think it through before you proceed.
- CORRECTION: Tim suggests that atom:content is limited to text and HTML. For clarification: valid atom:content @type values include "text", "html", "xhtml", and any valid MIME type. In the case of @type="xhtml", the content should be in the form of an XHTML fragment wrapped by an XHTML div.
11-10-2005 03:38:01PM - Permalink - Comment [1] - Trackback
category: XML
related topics: (Atom) (syndication)