I am working with Kevin Goff's book on VS 2005 Reporting and trying to learn "his" way of creating strongly typed datasets. He is using these dataset for Crystal Reports but I plan to use them in data entry forms as data source of controls. So I was creating some test methods to see what they return and so on. Therefore, the point of making certain methods static was simply for convenience of calling them.


Hi, Dmitry,

Just a clarification - I used the XML representation of typed datasets as the design-time source for CR reports, and then populated an instance of them at runtime, when I wanted to "push" the report data into the report.

What you're doing has at least a certain amount of commonality - you are using them as a design-time source for your data-bound controls.

(Both methods, in varying forms are fairly common).

As far as physically creating them to begin with...at least for reporting apps (which was the subject of the book)...since I often started with writing stored procs and testing stored procs, I wound up writing a utility that would execute each stored proc and output the results structure as an XSD. After that, I could create a new typed DS and point to the XSD (as opposed to rebuilding/re-typing the names). So basically every major .NET reporting system I've done has had a separate project with nothing but typed dataset definitions, used as references for any report that needed them.

Let me know if that helps...