Cart
[Home] [By Thread] [By Date] [Recent Entries]
>We need to develop a process for doing XML datamodelling, so that it can >be become a validated science and not just magical art. Then they even >might begin to teach this methodology at the universities, just as they >teach relational algebra and modelling. >capabilites to their existing database systems, making it very easy to >work with XML data/documents on the storage level. This is more or less my point. Our data modelling experience has generally been done under the influence, direct or indirect, of a well-understood relational paradigm. This is apparent in most schemas out there. I fully agree that we need to rigorously develop XML data modelling practicies, but without XML-based persistence solutions, this would be nothing more than an academic exercise. We also need a standardized XML DB interface (I know that XML:DB has done some work in this area)- W3C's query language is a start, but doesn't address update issues, to my knowledge. One other point about redundant data. If it can be done with relatively small footprint cost (each instance taking up only a small delta once the first instance has been stored), then the redundant solution is not necessarily as bad as it looks, particularly if the ability exists to change all instances in a single command from the application viewpoint. Plus, it gives the flexibility to, for example, change the customer address for POs that haven't been shipped, while retaining the address for those that have. Michael C. made a good point about not really knowing what RI even means in an XML context- back to data modelling, again! Other features that people have grown accustomed to in an RDBMS- triggers, stored procedures & the like- are more plumbing issues for the most part than they are necessarily intrinsic to the XML information model and will be available as products mature, although, again, we need a standards-based way of specifying/invoking these things. Linda
|
