feed parsing use cases

Feeds (e.g., RSS 1.0, RSS 2.0, Atom), are a real-world usage of mixing 
HTML and XML.

Two use cases, exemplified by Planet (Venus|Mars):

1) Taking a feed which was produced by a variety of means ranging from 
XSLT transforms to hand-coded templates into a single unified DOM which 
can be sanitized, scanned for relative URIs that need resolving, etc.

2) Taking a DOM and producing a feed which can be consumed by a range of 
consumers employing everything from "real" XML and HTML5 parsers to a 
variety of incompatible (tag|beautiful) soup parsers to regular expressions.

My best guess is that the general form of the solution to #1 is 
something like Anne's XML5 proposal though one modified to allow a much 
larger set of pre-defined entity definitions.

My best guess is that the general form of the solution to #2 is an XML 
serialization which avoids the use of namespace prefixed elements and 
empty element syntax, and uses well-known prefixes for popular namespaces.

- Sam Ruby

Received on Monday, 3 January 2011 14:02:51 UTC