Merge pull request #186 from dbrobins/scraper-abstraction

Scraper abstraction: provide an encapsulation.

To eventually allow scrapers to be written in Python or other languages, first the scraper interface itself (CScraper) was changed to define several language-agnostic methods (in particular, with no reference to XML): FindMovie, FindAlbum, etc., replacing the old Run (now internal).

Most XML parsing has been moved into CScraper, which for now functions as both a general scraper interface and the implementation for XML scrapers (in future may split into IScraper / CXMLScraper).

CScraper now ensures its scraper is loaded by itself; loading is an implementation detail that clients should not have to worry about.