I agree completely. The responsibility for making sure that the data is correct and well-formed is falls upon the people generating the data. I've already addressed this issue with user training and a filter to encode entities and before the data entry people mark up the articles.
Unfortunately, I inherited this project after about 400 articles had already been scanned and marked up.
----
Coyote
Comment on Re: (Ovid) Re: Dealing with Malformed XML