I envision having two revisions of a resource or a subtree thereof (filtered be a query in the first place) side by side and showing the differences. This has the huge benefit that we do not do line-by-line diffs. Instead we are really diffing the tree-structures and thanks to rolling hashes optionally stored fod each node the diffing algorithm can skip whole subtrees from the comparison :-) furthermore I might add a very simple index (which node keys haved changed and how) if you compare consecutive revisions.
I envision having two revisions of a resource or a subtree thereof (filtered be a query in the first place) side by side and showing the differences. This has the huge benefit that we do not do line-by-line diffs. Instead we are really diffing the tree-structures and thanks to rolling hashes optionally stored fod each node the diffing algorithm can skip whole subtrees from the comparison :-) furthermore I might add a very simple index (which node keys haved changed and how) if you compare consecutive revisions.