Hi everyone,
Some publishing houses have established URL templates that can be followed for scraping accurate high-quality images for covers. For instance, O'Reilly follows a specific standard for their images by ISBN:
http://akamaicovers.oreilly.com/imag...101848/lrg.jpg
For O'Reilly, ISBN-13 should work for anything in their catalog. I have also found some by ISBN-10, when ISBN-13 was not applicable.
I didn't notice anything like this being done in Calibre's current scrapers, so I thought I would put this out into the open as an idea for future development.