Update properties.json with reverse link names

[lzehl]

@apdavison I did not use the fairgraph suggestions, because I don't think that they are all correct (and avoiding ambiguous reverse linkages).

@apdavison @Raphael-Gazzotti @olinux I semi-automatically generated the ones in this PR now for each property (without over harmonizing which I'm not really sure is working several cases).

For now I would suggest to merge this (if there is no obvious error) and then we can revisit the whole concept and discuss about conventions (also for property names in general) in the future.

[apdavison]

I don't see any obvious errors, although it is difficult to review because if I expand the context my browser hangs due to the size of the file, so let's merge this and fix any problems as we encounter them.

[Raphael-Gazzotti]

General comment: I’m not sure we need reverse links to QuantitativeValue or QuantitativeValueArray instances, since identical measurements would typically be reused rather than linked separately. That also feels a bit inconsistent with our perception of embedded types.

[lzehl]

@apdavison @olinux @Raphael-Gazzotti for the QuantitativeValue or QuantitativeValueArray instances or actually all embedded schemas: do we need reverse links for those or should we explicitly not provide them? (may also be influenced by the decision on are schemas either always linked or always embedded?)