Lists Home |
Date Index |
The statement was motivated by an earlier thread on this list, where one person was obviously thinking in XML document terms, and the other in XML database terms. I found that thread frustrating because the two people in it were obviously talking at cross-purposes- an XML database is not simply a collection of documents, which is perhaps a better way of stating it.
Moreover, there were (are?) several XML database offerings that used relational engines under the hood, so there is a distinction to be made between those implementations and the ones that started "from scratch" with the underlying assumption that XML would be the data stored. Both called themselves "native", although one was a wrappering of relational technology. And if you can succeed in wrappering it efficiently enough, eliminate schema dependencies, and provide a pure XML interface, who's to know the difference?
I also wonder- if you have a complex set of related XML documents, such that different document types point to one another in relationships that are not well-modeled hierarchically, do you end up building something that looks pretty relational anyway? If you want to avoid lots of data duplication, you probably do. Hence the concept of the XML-Relational hybrid, I suppose. Don't know if it really works, but I can see where it might come in handy.
From: Michael Kay <email@example.com>
Sent: Aug 25, 2004 10:41 AM
To: 'Linda Grimaldi' <firstname.lastname@example.org>,
'Owen Walcher' <email@example.com>, firstname.lastname@example.org
Subject: RE: [xml-dev] are native XML databases needed?
> At least with the implementations I am familiar with, natvie
> XML dbs do their own version of shredding. Randomly
> accessing XML document contents efficiently requires this.
> Now, their implementations are optimized for XML, an
> important difference from relational approaches, but their
> storage image is far from document form.
I think it's a little far-fetched to treat as equivalent a system with an
internal data structure designed specifically as an implementation of the
hierarchical XML data model, and a system that constructs a tabular
representation using a data model that was never even designed to enable
recursive queries. If you've ever had a corridor conversation with someone
trying to implement mixed content or comments and PIs or namespaces in a
shredded table, you can only feel sorry for the guy.