Last modified: 2009-06-22 23:35:17 UTC
Not sure if this is not yet done because of any compatibility concerns/issues, but: Consider using CDATA or <summary/content type="xhtml" xml:base="http://en.wikipedia.org/"><div xmlns="http://www.w3.org/1999/xhtml"> for Atom feeds instead of escaping all those <, >, " and maybe also some &. See http://de.wikipedia.org/wiki/Atom_(Format)#Unterschiedliche_Inhaltstypen Would reduce size/traffic a bit, but like I wrote: Not sure whether this might be a client issue or not.
CDATA would work for RSS too, see http://webservices.xml.com/pub/a/ws/2002/11/19/rssfeedquality.html (bottom, tested)
Think I'd prefer not to do this... IMO use of CDATA makes the resulting code harder to read since the meta-level boundaries become less clear, and would make only modest changes to the output size. XML subcontent for Atom feeds also requires 100% pure correctly valid (or at least well-formed?) XHTML output which I'm not sure how much we can guarantee, or else the feed breaks. Poorly written RSS consumer code is also fairly likely to get confused by CDATA, for instance where trying to use regexes instead of a proper XML parser to fetch desired bits of data.
