The RSS for the Events PI often contains HTML characters in the <description> that have to be escaped somehow to be considered valid XML. The Events PI has a toggle switch to indicate whether it should expect HTML in the description element.

Either individual HTML characters can be replaced with entity references e.g. < replaced by & lt ; and > replaced by & gt ; or alternatively you can enclose the HTML in a CDATA section e.g. <![CDATA[ <b>Our event!</b> ]]>.

Both SharePoint and Drupal use CDATA sections to enclose HTML in the RSS views they generate by default, rather than escaping the individual characters, which seems more efficient and is definitely valid XML. Unfortunately the Events PI will literally display the HTML code if the RSS includes CDATA sections leading to the somewhat unfortunate HTML 'rich' events as seen in the example attachment. The only work around is to intercept and transform the HTML before it is ingested by the Events PI, which is yet another thing to maintain.

Ex Libris Support asked me to raise this feature request as support for CDATA is not currently planned, but given that big platforms generate RSS in this format it seems to me like it would be a 'good thing', and arguably what one might expect to happen anyway. Thanks in advance for the up votes!