Evidence before optimization

Performance recommendations should map to a measured bottleneck. We favor controlled changes, representative testing, and rollback plans.

Product claims

We use vendor documentation for current product capabilities and avoid presenting marketing claims as independent benchmark results. Comparisons should identify architecture and use-case differences rather than manufacture universal winners.

Corrections and updates

Software changes. We review commercial pages for material feature changes and update statements that are no longer accurate.