Roger Costello wrote: An XML Schema defines a new syntax
Ghislain Fourny responded: Validation is done on top of the data On the other hand, schemas that leverage Thanks for the feedback Ghislain! I probably have an incorrect understanding of what a “data model” is. I think of a data model as a way of viewing the world. For example, the XSLT and XQuery
folks view the world as a bunch of “sequences”. For example, “this element consists of a sequence of items.” That model is different from other models such as the Infoset model or the PSVI model. But the “sequences” concept is very useful to the XSLT and XQuery
folks. The XSLT and XQuery folks could have used an existing model, such as the Infoset model, but they chose not to. I suspect that they devised a new data model (XDM) because existing ones (such as the Infoset) are too general; they needed something more
specific. Similarly, when I begin work on some data, I want all the users of the data to have a specific, agreed-to view of the data. I could use an existing data model
such as Infoset or PSVI, but they are typically too general. That’s why I think it would be good to create a data model prior to creating any XML Schema. What do you think? /Roger |