This is ultimately the discussion we should be having, which is whether using Schema+Namespaces is preferrable to using straight serialized data (RSS-Data/SDL). Here's a running experiment from Les Orchard which is comparing the two approaches:
Right now, he's simply shown what RSS with namespace'd Amazon data looks like inline to an RSS item, compared to inlined RSS-Data. Next, presumably, he'll give us a look at how much code is required to parse and use the data, which is part of the issue here.
Clearly, namespaces with schema provide the benefits of a human readable format for the given type of data, and a contract that can be consulted on its data structure. The downside is that it is a bit more work (we'll see) for developers to consume and use that data.
In a private thread with a close friend and colleague (Sean Neville), he expressed the view that the semantic richness of Schemas are worth the pain in parsing/consuming, and that furthermore the tools and libraries to handle schematized data were getting abstract enough to make it a moot point:
I suppose if you're handcoding or manually parsing data as XML that namespaces get cumbersome, but considering the fact that tools across all platforms, and especially those debuting in MS Office for line of business folks, abstract and employ schemas for content typing, I can't imagine this occasional complexity for a small group of developers (small relative to the number of end users who stand to benefit from schema's data richness) is a reason for doing away with it altogether. The flexibility and power of the format, combined with the abstracting away of the complexity, makes this dealable. Whether an element is harder for some RSS developer to parse manually isn't as relevant to me as is the richness of the content and simplicity to end user subscribers. Generally speaking, I think the approach of making schema simpler to use for a larger group of people is preferable to replacing it altogether with something that's likely less rich and only simpler for a short time for a small group of folks.
I'm interested to hear how many RSS/XML developers out there currently rely on XML Schema for their XML documents? How do you find the tools, and how would you compare that to the simplicity of simple data structure exchange via something like RSS-Data?