PORTAL BENCHMARK — BLANK OBSERVATION WORKSHEET
No measured timings supplied. Synthetic fixtures, not production sizing advice.
Pro version: ____ OS: ____ Host/version: ____ Network: ____
Layout/occurrence: ____ Window/rows: ____ Dataset: small / large
Task: switch P02 to P01; end when visible rows rendered and usable.
Method/resolution: ____ First-visit observation: ____
First visit is not proof of empty client/server/OS caches.
Variant: baseline / sort / Open filter / named object-calculation / volume
Exact change: ____ Expected IDs/order: ____ Actual correctness: ____
Five warmed runs (seconds): ____ ____ ____ ____ ____
Median: ____ Minimum: ____ Maximum: ____
Repeat baseline/candidate in alternating order with same endpoints.
Small: 100 tasks per project, 200 total, 50 Open per project.
Large: 1000 per project, 2000 total, 500 Open per project.
Even Sequence=Open. Odd=Closed. No cross-project membership.
Use separate copies for small/large; never import both into one baseline.
Check updates, empty parent, access rules and totals separately.
Conclusion: consistent / inconclusive / regression. Reason: ____
Do not generalize local timing to WAN, Go or WebDirect.