Event Data mapping domains to DOIs gives false precision but not much useful accuracy

Where?

As documented in the landinge page flow chart, the Percolator includes lots of information about how it tried to validate the mapping of landing page to DOI.

What's the situation?

The Percolator doesn't trust a publisher domain to represent its DOI. It goes through a number of steps to verify and match, recording each one.

This is closely related to #93

What does it make more difficult?

There's a lot of precise data here, but it's difficult to interpret. It means that the Percolator has to do a lot more work (slowing down throughput). The success is also variable and can be affected by external factors (such as network issues) which may bring in more confusion.

How can we improve it?

Radically simplify. Keep only a list of domains. Trust those domains to correctly represent their DOIs.