If you were to limit the revision deletion to groups of revisions by a single author within a certain timeframe (a week? a month?), would it lose most of the size savings? It would be nice, IMO, to keep a longer-term history of an article, while squashing a flurry of small changes into one larger one.
But also, saving the diffs instead, why would that break things? The deletion of a revision would just result in the squashing of the two diffs, right?
If you were to limit the revision deletion to groups of revisions by a single author within a certain timeframe (a week? a month?), would it lose most of the size savings? It would be nice, IMO, to keep a longer-term history of an article, while squashing a flurry of small changes into one larger one.
But also, saving the diffs instead, why would that break things? The deletion of a revision would just result in the squashing of the two diffs, right?