[JoJo Gunn]

Test with other well known links:

http://www.drudgereport.com/

http://www.foxnews.com/

http://www.washtimes.com/

http://groups.yahoo.com/group/freerepublic2/

Looks good in the preview....

[John Robinson]

It was the ampersand. To get technical, I forgot # is a comment character within a regex when under the influence of Perl's /x modifier.

m{ <\w+[^>]+> | &\#\d+; | &\w+; }x

(Yes, that actually means something. It returns true if the target string contains anything resembling HTML.)