You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Wasserförderung: review of feat/wasserfoerderung-plan-a and the way forward
This reviews our own branch against FI 5.008 (Staatliche Feuerwehrschulen Bayern), the Wasserförderung guidance a Lehrgang teaches. The branch is #152 plus #164, which was merged into the branch and never reached main; a build of it caused #396. Below: what is wrong with the branch, what it gets right, what the feature is missing, and how to get it shipped.
The FI 5.008 figures quoted here (Tab. 3 friction, the Abb. 22/23 worked examples) still need checking against the document itself before #584 builds golden tests on them.
1. What is wrong with our approach
It never shipped. One PR grew into a calculator, a hand-written map control, a custom elevation format ("FWDM"), a region-pack download pipeline, sync and PDF.
It has been open for six weeks and now conflicts with main.
Five review findings are still parked: an unobserved download fault, an MBTiles connection per tile, duplicate factories.
Our friction table looks optimistic. We use 1,0 bar/100 m at 800 l/min. FI 5.008 Tab. 3 gives 0,24 bar per 20 m, which is 1,2 bar/100 m:
l/min
ours (bar/100 m)
FI 5.008 (bar/100 m)
600
0,6
0,7
800
1,0
1,2
1000
1,5
1,7
1200
2,25
2,4
At 8 bar out and 1,5 bar in, FI 5.008 gives a 540 m flat stretch (Abb. 22) where we plan 620 m. Our pumps end up about 15 % too far apart, which is the unsafe direction. The doc comment calls our table "verified" but cites no source.
Downhill has no limit. On a descent Cost() goes negative, so a stretch keeps getting longer. Nothing checks the resulting pressure against the hose or the pump; a B hose is rated for 16 bar.
The output is a count, not a plan. A Leitung stores PumpCount, ReservePumpCount and pump positions in metres. It stores no Eingangsdruck or Ausgangsdruck per pump, no rise per section, no chart and no vehicle. The Kommandant learns "3 Verstärkerpumpen" but has nothing to hand to anyone.
The map doesn't show the pumps.MapDrawing draws the route but not the pump positions we compute along it.
Plan A is one length plus one rise. That assumes an even gradient and hides every crest. The only crest-aware path needs a downloaded region pack.
The PDF footer is wrong. It hardcodes "B-800, 8 bar Speisedruck … 3 % Reserveschlauch", although every Leitung stores its own FlowLMin and FeedPressureBar. The 3 % is also not spare hose: it is a pressure margin. Reserve hoses are ceil(L/100).
"3 % of the budget" is a unitless setting that nobody learns on a Lehrgang. FI 5.008 expresses the safety reserve in bar.
What the branch gets right:
It works offline, with region packs and a local elevation model.
The plan is stored and never recomputed, so the PDF and sync can't drift. That matches the "values, not references" rule.
An incident can hold several Leitungen.
The crest check uses the worst point inside a stretch, not just its ends.
The plan is part of the Bericht.
2. What the feature is missing
Addition
Why
Druckverlauf chart: terrain, pressure sawtooth, min. Eingangsdruck line, numbered pumps
Explains the plan at a glance; in the UI and the PDF
Result per pump: role (Wasserentnahmestelle / Verstärkerpumpe / Einsatzstellenpumpe), station, Höhe, Eingang, Ausgang, section length, rise and B-20 count
Turns a count into an order
Einsatzauftrag per pump position: Funkrufname, UTMREF and WGS84, pressures, hoses to the next pump, neighbours up- and downstream, Rückmeldung
Online routing, geocoding or elevation. They send the Einsatz location to third parties and fail without LTE.
what3words.
Plus Codes as the primary location.
Base maps for Bavaria only.
A DBV setting value, until it has a source.
3. The way forward
Stop rebasing #152. Keep the branch as a source to copy from, and land the work on main in small slices. Each slice renumbers its own migration at merge time.
Activity
feat(wasserfoerderung): downloadable region packs for Einsatzgebiet (#…
fix(wasserfoerderung): center Karte mode on the configured region's r…
feat(wasserfoerderung): draw Förderstrecken on a map with real elevat…
feat(wasserfoerderung): Förderstrecken planen — Rechner, Tab, PDF (#150)
feat(wasserfoerderung): draw Förderstrecken on a map with real elevat…
feat(wasserfoerderung): Förderstrecken planen — Rechner, Tab, PDF (#150)
feat(wasserfoerderung): draw Förderstrecken on a map with real elevat…
DeepDiver1975 commentedon Oct 15, 2026
Wasserförderung: review of
feat/wasserfoerderung-plan-aand the way forwardThis reviews our own branch against FI 5.008 (Staatliche Feuerwehrschulen Bayern), the Wasserförderung guidance a Lehrgang teaches. The branch is #152 plus #164, which was merged into the branch and never reached
main; a build of it caused #396. Below: what is wrong with the branch, what it gets right, what the feature is missing, and how to get it shipped.Slices are tracked as:
1. What is wrong with our approach
It never shipped. One PR grew into a calculator, a hand-written map control, a custom elevation format ("FWDM"), a region-pack download pipeline, sync and PDF.
main.The scope is what stopped it.
Our friction table looks optimistic. We use 1,0 bar/100 m at 800 l/min. FI 5.008 Tab. 3 gives 0,24 bar per 20 m, which is 1,2 bar/100 m:
At 8 bar out and 1,5 bar in, FI 5.008 gives a 540 m flat stretch (Abb. 22) where we plan 620 m. Our pumps end up about 15 % too far apart, which is the unsafe direction. The doc comment calls our table "verified" but cites no source.
Downhill has no limit. On a descent
Cost()goes negative, so a stretch keeps getting longer. Nothing checks the resulting pressure against the hose or the pump; a B hose is rated for 16 bar.The output is a count, not a plan. A Leitung stores
PumpCount,ReservePumpCountand pump positions in metres. It stores no Eingangsdruck or Ausgangsdruck per pump, no rise per section, no chart and no vehicle. The Kommandant learns "3 Verstärkerpumpen" but has nothing to hand to anyone.The map doesn't show the pumps.
MapDrawingdraws the route but not the pump positions we compute along it.Plan A is one length plus one rise. That assumes an even gradient and hides every crest. The only crest-aware path needs a downloaded region pack.
The PDF footer is wrong. It hardcodes "B-800, 8 bar Speisedruck … 3 % Reserveschlauch", although every Leitung stores its own
FlowLMinandFeedPressureBar. The 3 % is also not spare hose: it is a pressure margin. Reserve hoses areceil(L/100)."3 % of the budget" is a unitless setting that nobody learns on a Lehrgang. FI 5.008 expresses the safety reserve in bar.
What the branch gets right:
2. What the feature is missing
Deliberately not planned:
3. The way forward
Stop rebasing #152. Keep the branch as a source to copy from, and land the work on
mainin small slices. Each slice renumbers its own migration at merge time.feat(wasserfoerderung): FI 5.008 planner engine with per-pump pressures #584: planner engine (Domain only, no UI). Rework
FörderstreckePlannerto FI 5.008:Golden tests reproduce FI 5.008 Abb. 22 and 23. Reuse
ElevationProfileSample,GeoPointand the branch's crest logic.feat(wasserfoerderung): tab, persistence, sync and PDF with Druckverlauf chart #585: tab, persistence, sync and PDF.
MinimumProtocolVersion, plus aSchemaGuardentry.feat(wasserfoerderung): Einsatzauftrag per pump position with Funkrufname and UTMREF #586: Einsatzauftrag.
feat(wasserfoerderung): map mode salvage from #164 with pump markers and GeoJSON/GPX export #587: map salvage from feat(wasserfoerderung): draw Förderstrecken on a map with real elevation profiles (#150, Plan B) #164.
feat: wasserförderung und langewegstrecken - execution mode #87 execution mode builds on feat(wasserfoerderung): Einsatzauftrag per pump position with Funkrufname and UTMREF #586 and feat(wasserfoerderung): map mode salvage from #164 with pump markers and GeoJSON/GPX export #587.
Open before #584 starts:
🤖 Generated with Claude Code