As Josh mentioned in a recent post, SQL/MED makes it more likely that Postgres will be playing alongside lots of different data sources. This has been the story of most of my IT career, really. I have tended to operate at the borders between systems and make them play nicely together, rather than within the bounds of a single system or data source.
Yesterday, I got another Foreign Data Wrapper working. I'll be announcing what it is at my talk on FDWs at 1.30 pm on Thursday, so if you want to find out what it is come along This FDW is more self contained than the file_fdw example in contrib, or the text_array extension I've been working on, both of which use the new COPY API to do a lot of their heavy lifting. I have another FDW in the works too, which I'll also be talking about a bit on Thursday.
The talk description says I'll be talking about C coding, and there will be a bit of that, but it's really more about how to use FDWs than about how to create them. You don't need to be a code cutter to attend the talk!