The RSS Blog

News and commentary from the RSS and OPML community.

Is Nokia Lifeblog using the AtomAPI? Well, yes and no. First, let me point you to the draft specification. Then, let's pray that this feedback gets to them before they get out of draft? My notes follow...

  • The entire bloglist GET request is not part of AtomAPI. This means Nokia is both supporting the AtomAPI and creating a proprietary protocol that can't be re-used elsewhere, a.k.a. vendor lock-in.
  • The POST request in section 2.4 has an Accept HTTP header of text/xml. Wrong. Shouldn't this be application/atom+xml?
  • The POST request in section 2.4 has a Content-Type HTTP header of application/xml. Shouldn't this be application/atom+xml?
  • The POST request in section 2.4 is confusing. Make certain to note there are soft linebreaks in there that shouldn't be there.
  • The POST request in section 2.4 has an empty title/@type. The type attribute must be a valid media type.
  • The POST request in section 2.4 has a generator element with a missing required version attribute.
  • The POST response in section 2.4 has the Transfer-Encoding HTTP header of chunked. But the HTTP message does not use chunked transfer encoding.

Some of these points are explained away in other sections of the document. But this is simply not right. What they've created is a completely proprietary version of AtomAPI that could never be re-used. This is what happens when you introduce a half-baked protocol and leave it stagnant for almost a year. I wish the Atom working group would either get to work and finish it, or obsolete it, so we don't end up with any more horror stories (every AtomAPI implementation to date).

Reader Comments Subscribe
Type "339":