As I mentioned previously, I find myself conflicted about having my blog posts live in a database. And, really, I don’t see that much difference between a blog post and a generic XML file. (As I mentioned, I wish I could write posts in XML.) I feel that blog posts, like generic XML, don’t map naturally to a relational database, particularly if you want to have fine-grained access to individual elements. Matt Liotta and Chris Preimesberger discuss the possible performance problems you might run into by trying to store XML in an RDBMS, and how a more elegant solution lies in native XML databases that can be queried in more natural (at least for XML) XPath and XQuery instead of SQL. As the name implies, XPath (which XQuery utilizes) has a lot in common with file-system paths. Consider that the browser’s location field is better suited to handling a file-system path than a query written in SQL (and file-system paths are in fact how most blogs are queried—whether by date or category, regardless of whether the blog engine stores posts on the filesystem or in a database.) And, especially in a shared-hosting situation, I don’t know if a database really gets you all that much more performance than simply dealing with the file-system. Then again, considering that I don’t find hierarchical categories all that useful, I don’t know if paths are all that great either, except for accessing specific elements in an XML document. Decisions, decisions.