For the revision slider it could be useful to know what is actually done when comparing revisions. Thus I/we could optimize the design for the most common tasks instead of possibly going into a direction that does not cover frequent usecases.
- How many steps back in time do users go when comparing versions? (measured in current state -n revisions)
- How big is the difference between compared versions (measured in revisions)
- Is it common to jump between versions back and forth in a short time or do users find the revisions quickly without iterating and adjusting often?
All data makes the most sense if given in[[ https://en.wikipedia.org/wiki/Quantile#Specialized_quantiles | Percentiles/Deciles ]], so we can see the distribution.
Could be delivered as Excel/Calc, csv, RDataFrame
On a diff you have a "previous" and "next" edit button.
Would tracking how often these were used be of use?
Yes, that is a reasonable proxy for what interests me, I suppose.