Andrew Gray wrote:
So what we should be considering is some way of identifying these "risky" articles and doing something about them. Perhaps the ones who are of the most risk of hatchetry plus the least general importance should be prioritised, if we can figure out an evaluation system.
More could probably be done at a technical level, but given the difficulty of getting stable version implemented as a simple yes/no feature hope for something more sophisticated seems far more distant. I hate to criticize our technical people on this, because they must have their hands full just keeping up with the priority of making sure the whole enterprise keeps running.
I find it easy to suggest statistical evaluation processes that would do this. Perhaps it's just a matter of having someone who can devote himself to such initiatives without the obligations connected with daily maintenance to keep him distracted.
Ec