I noticed my feed elements have character entities (for fancy punctuation characters) within content marked as
CDATA, e.g. (in Atom feed)
<summary type="text/plain" mode="escaped"><![CDATA[Here is an apostrophe: ’]]></summary>
Is this normal?
My understanding was that all CDATA content is interepreted "as is" (no markup, no entities) -- so charcater entities would not be decoded.