I could have titled this post as "lesson from the field". In this blog post, "How XML Threatens Big Data" is a report about experience made with using XML for some data projects, like biotech, and then continues to give "Three Reasons Why XML Fails for Big Data" and then "Three Rules for XML Rebels".
I was pulled to some IT projects where the focus was on doing everything in XML. People got overboard in adopting new technology. However, DB2 didn't scrape its relational capabilities for a reason and supports both relational and XML data side-by-side. This is because there a projects where relational format (a.k.a. tables with columns and rows) is the right storage format and there are projects where it is XML. In many cases both are needed and it is good that DB2 is a hybrid system, supporting relational data and XML data very well.