esr speaks it:
All data, including primary un-”corrected” datasets, must be available for auditing by third parties. All modeling code must be published. The assumptions made in data reduction and smoothing must be an explicitly documented part of the work product.
These requirements would kill off AGW alarmism as surely as a bullet through the head.
Transparency would kill off AGW denialism, too.