Seems Clear Enough from the Specification

» Seems Clear Enough from the Specification #

Paul R. Brown @ 2008-08-12

I can understand how people get confused with RSS since what passes for a specification isn't exactly precise, but the Atom syndication spec (a.k.a. RFC 4287) is nice work and comes with an unambiguous RelaxNG grammar. Which is why I find this (sent by a friend who reads this blog in Google Reader) irritating:

mis-rendered title

This is wrong. (For what it's worth, this doesn't happen for me in Google Reader running in Firefox 3.0.1.) Extracting a bit of the grammar, the title of the feed is an atom:title element, and XHTML content is supported:

atomTitle = element atom:title { atomTextConstruct }

anyXHTML = element xhtml:* {
   (attribute * { text }
    | text
    | anyXHTML)*
}

atomPlainTextConstruct =
   atomCommonAttributes,
   attribute type { "text" | "html" }?,
   text

atomXHTMLTextConstruct =
   atomCommonAttributes,
   attribute type { "xhtml" },
   xhtmlDiv

atomTextConstruct = atomPlainTextConstruct | atomXHTMLTextConstruct

xhtmlDiv = element xhtml:div {
   (attribute * { text }
    | text
    | anyXHTML)*
}

So the construct from my feed is valid:

<title type="xhtml">
  <div xmlns="http://www.w3.org/1999/xhtml">mult.ifario.us - All Posts</div>
</title>

I spent time with the excellent and useful Feed Validator when I was writing perpubplat, even compromising on things that I felt I shouldn't have to compromise on, e.g., namespace prefixes, in the name of interoperability. The validator awards the feed a squeaky clean bill of health. Nonetheless, this issue makes my feed ugly in Google Reader, since the UI shows an abbreviation like <div xmlns="... in the sidebar and elsewhere. (FWIW, this Google Reader issue isn't caught by the title conformance tests.)

I've posted a bug in the relevant Google Group here.

prb.io

» Seems Clear Enough from the Specification #