Allow the creation of multiple architectures for one architecture in the paper, for each researcher

[nicoSix]

In many SLRs, a paper might be read by multiple researchers and merged later to avoid bias.
It would be nice to allow creating multiple features to support that:

• Disable architectures once they are "merged" => will not count in statistics and pattern inference
• Get the divergence ratio between two architectures to know the level of difference there is between them (maybe something like Kappa Cohen ratio?)

This is not trivial and requires a bit of research to be implemented efficiently