Berkeley DB has been the standard for low-profile and fast UNIX database needs for a long time. Development has continued all the way until today. For a couple of years, most of the work was done by Sleepycat Software while the product is still open source.

DB XML is an extension of the DB system bringing complete XML support to the database (using Xerces and Pathan for XML and XPath parsing). This opens the oppurtunity for applications to stop squeezing the hierarchical, flexible XML into arbitrary, developer-defined flat table structures and use the full power of XML. I strongly believe, that XML databases are soon to take off once the teething problems have been worked out and developer knowledge on critical technologies like XML and XPath has increased.

XPath is the query language of choice for accessing XML data with the new product. Unfortunately, the atom of Berkeley DB XML is a XML document and not a single XML element. This means, XML data can be added and replaced only as a complete set of nested elements, while retrieval (due to XPath) is possible on an element level.

While this solution adds a nice feature to the database system it is still not sufficient for a complete XML based solution as a full-grown application would need to update, insert and delete single elements, attributes or text nodes on demand on any node of the XML document. But I am not surprised as the necessary standards are not here yet. XPath 2.0 will be enhanced significantly in order to support an extended way of addressing data and will be the basis for XQuery 1.0 which is positioned to become the SQL of XML in the future. [via Notes From My Terminal]