I'm writing a windows service that processes a number of different rss news feeds at regular intervals. These news items will be saved into our database and associated with different objects in the system.
Although there is a set specification on what needs to happen, there is no UI component for the customer to verify.
What's the best way to write acceptance tests for something like this?
Should I create some simple web pages that display a summary of data that needs to be verified?
Since the data is stored into a database the customer can verify it by reading the database with an IDE or by dumping the data to excel/csv.
I would recommend against doing a lot of extra work to make it possible for them to verify the results because they may end up testing the verification procedure more than the real underlying program.
For internal testing, we often rely on logging for testing. We tell testers what logs to look for for good/bad results.