Chapter 39 — Exercises
Work these with the chapter closed where you can. Where a question touches a regulatory requirement — drone operation especially — answer by function ("somebody has to be certified, and here is how I would find out what that means where I work") rather than by rule number. A manager who quotes an aviation rule number out of a textbook will be wrong somewhere and out of date eventually.
Difficulty legend: ⭐ basic · ⭐⭐ applied · ⭐⭐⭐ advanced judgment · ⭐⭐⭐⭐ extension and research
Selected answers appear in Appendix J. Calculation items have their numeric answers in a <details> block here — try them before you open one.
Part A — Conceptual Understanding ⭐
A1. Name the four failure modes from §39.1 and define each in one line. For each, give the sentence you would expect to hear in a sales presentation that signals it.
A2. §39.1 argues the four failure modes are "four faces of one economic fact." State that fact in one sentence, then explain why it bites construction harder than it bites manufacturing.
A3. The chapter offers a single screening rule that "explains almost everything in §39.2." State it, then apply it to three technologies in the chapter and say whether each passes.
A4. List the five gate questions from §39.9.1. Which one does a vendor find hardest to answer, and which one does your own organization find hardest? They are not the same question, and the difference is the point.
A5. Explain the rule "D is a gate, not a score." Then explain how a proposal can score 17 out of 20 and still deserve to be killed.
A6. A service provider quotes $2,600 to scan an existing mechanical penthouse and your project executive concludes you now have a model of it for $2,600. Correct him using the chapter's rule of thumb about capture time versus modeling time. Then answer a second question: why does §39.2.3 insist the reality-capture loop must be short, and what does the same activity become when the loop is long?
A7. Describe the five rungs of the prefabrication ladder in your own words. Which rung does virtually every commercial contractor already use without calling it prefabrication, and which rung are people actually arguing about when they say "we should do more prefab"?
A8. Restate in your own words, with an example from your own experience of any kind: "A platform does not create discipline. It makes the absence of discipline visible faster."
A9. State the chapter's position on robotics and construction labor in two sentences — the version you would say to a reporter, a worried apprentice, and a parent at a career day, without softening or overselling it in either direction. Then give three examples of a technology that removes work and three that moves work.
Part B — Applied Analysis ⭐⭐
B1. For each of these five sentences from real-shaped vendor materials, name the failure mode it signals and the question you would ask next.
- (a) "It works best when the whole project team is on the platform."
- (b) "It only takes a few minutes a day from each foreman."
- (c) "You'll have complete visibility into field productivity across your portfolio."
- (d) "In a controlled setting it achieves a consistent cycle time."
- (e) "Onboarding your subcontractors is free."
B2. A specialty contractor offers to run your overhead coordination and give you a fabrication-level model at no additional cost — on the condition that they fabricate the racks in their shop and you commit the scope to them before the design is complete. Score it on A/B/C/D, name what it breaks, and write the two contract terms you would insist on before saying yes.
B3. Your team flies a drone monthly and produces a photogrammetric surface each time. Over three flights the computed cut volume has been erratic — up, then down, then up — while the work proceeded steadily. Using §39.3.2, list five candidate causes in the order you would investigate them and say what evidence would confirm or eliminate each.
B4. Your mechanical subcontractor has $1.4 million of prefabricated racks in his shop and has begun asking, politely, when he gets paid. Your prime contract permits payment for materials stored off site only with the owner's specific written consent. Walk through what you must assemble, whose signature you need on each item, and what happens to Kestrel's cash if the owner simply declines. Note where the answer depends on jurisdiction and on the specific contract form.
B5. An artificial-intelligence submittal review tool compares a product data sheet against the specification and returns "no exceptions." Six weeks later the product arrives without the required fire rating. Who is contractually accountable, what does the tool's terms of service change about that answer, and what do you change in your process on Monday?
B6. Kestrel has scanned eleven scopes in a year. Nobody has been assigned to review the deviation maps; they sit on the project server, well organized and correctly named. Run the four failure modes on the program as it is actually being operated. Which one has it drifted into, what is it worth in this condition, and what is the cheapest possible fix?
B7. Curtis Boone, running Rivermont Elementary School #12, buys the same field management platform Kestrel uses and tells his owner the project has "full digital document control." Six months later the RFI log holds 34 open items, 19 with no ball-in-court, and no aging report has ever been opened. Using Chapter 25, explain what he bought and what he did not. Is the platform net positive or net negative for him in a dispute?
B8. A 360-degree walkthrough program is running at 62% coverage. Argue that a partial record is worse than no record, then argue it is obviously better. Which do you act on, and what threshold would you set?
Part C — Calculations and Deliverables ⭐⭐–⭐⭐⭐
C1. Score the quarter and recommend ⭐⭐⭐
You are Grace Lindqvist. Three proposals, ninety minutes, and Nadia Haddad wants a recommendation on each.
| Proposal | Cost | Adoption friction | Who pays / who benefits | |
|---|---|---|---|---|
| A | Sell the as-built model as a deliverable — registered point cloud plus modeled overhead MEP in designated areas, added to Kestrel's CM at Risk proposals | $66,000 to the owner; $44,000 to deliver (outside modeling); $18,000 internal setup for a scope template and a written accuracy-and-reliance statement | Kestrel alone | Owner pays; owner's facility group benefits on every future renovation; Kestrel earns a margin and stops giving a product away |
| B | Subcontractor daily-report and manpower module — all 14 subcontractors log daily manpower and installed quantities in Kestrel's platform | $47,000/year in subcontractor seats plus Kestrel administration | Every subcontractor's foreman types daily. None of them bought the platform | Kestrel pays; Kestrel's office reads the dashboard |
| C | Passive arm-support exoskeletons — 40 devices for Kestrel's self-perform general-trades carpenters doing sustained overhead blocking, backing, and ceiling work | $52,000 one-time plus $6,000/year fitting, replacement, and training | Kestrel's own craft workforce; voluntary wear; fit varies by person | Kestrel pays; Kestrel's own workers and Kestrel's own injury experience |
Tasks. (1) Score each on A/B/C/D. (2) Recommend fund, kill, or restructure, in three sentences each. (3) One of these does not fit dimension D cleanly. Identify it, explain why, and say what you would do about it.
Worked answer
| Proposal | A | B | C | D | Total /20 |
|---|---|---|---|---|---|
| A — As-built as a deliverable | 5 | 5 | 4 | 4 | 18 |
| B — Subcontractor module, as proposed | 2 | 1 | 1 | 2 | 6 |
| C — Exoskeletons | 5 | 5 | 5 | — | see below |
A — fund it. The owner pays, the owner benefits, and Kestrel earns $22,000 per sale ($66,000 − $44,000) on a capability it already has and has been giving away. Do one thing before the first sale: write the accuracy-and-reliance statement — what the model represents, what its accuracy is, and what it is not (not a design document, not a warranty of concealed conditions, and it does not show what the scanner could not see). Reliance is where liability lives, and you are creating a document a facilities engineer will rely on for twenty years.
B — restructure, do not simply kill. As written it is a 6: the Simultaneity Trap (one holdout makes the dashboard silently wrong) stacked on the Displaced Burden (subs type, Kestrel reads). But the underlying want is legitimate. The §39.1 fix is to make the entry a byproduct of something the subcontractor already has to do to get paid — manpower tied to the daily report the subcontract already requires, quantities tied to the pay application. That lifts B and C to 3. Then answer D honestly: if the answer is "Wei Chen adjusts the six-week look-ahead when a trade's manpower falls below its committed level three days running," you have a real D. If it is "we'd have better visibility," you have a dashboard. And if you cannot make it a byproduct, price it in the subcontract instead of pretending it is free.
C — fund it, and notice the framework bending. A, B, and C all score 5. Dimension D does not apply cleanly, because an exoskeleton is a control, not a report. Do not fudge a score. Promote the question to gate 5, where §39.9.2 says physical and repeatability questions belong: is there a specific sustained-posture problem in a named scope, is the fit right for each wearer, will workers accept it, and — the one people forget — does the device create a new hazard such as a snag point, restricted movement, or impeded egress? Answer those four in writing and fund it. Otherwise you are buying forty devices that will live in a job box.
The general lesson: A/B/C/D was built to screen information technologies, which fail most often. Point it at a physical control or a construction method and expect D to bend and gate 5 to carry the weight.
C2. Design the pilot — with a kill criterion ⭐⭐
Take Proposal C forward. Design the pilot using all six elements from §39.9.4, in writing, before any device is bought. Deliver one table with six rows: scope, named owner and executive sponsor, window, pre-registered success criterion in numbers, cost cap, kill criterion.
Two constraints. Your success criterion may not be an injury statistic — you will not have enough events in a ten-week pilot on one crew to measure anything, and a criterion you cannot measure is a criterion you will argue about later. Your kill criterion must be an observation, not a mood — something a person could point at and end the pilot the same afternoon.
A model answer — yours may differ, but it must be this specific
| Element | Specification |
|---|---|
| Scope | Overhead blocking and backing, level 2 only, one project. Kestrel self-perform general trades: 12 carpenters, 8 devices rotated. Nothing else, nowhere else |
| Owners | Dani Okonkwo day to day (fit, logistics, data); Bea Salgado technical; Nadia Haddad executive sponsor, because a voluntary-wear program dies the first time a foreman turns it into a requirement, and only an executive can prevent that |
| Window | 10 weeks, fixed start and end date |
| Success criterion (pre-registered) | (a) By week 6, at least 70% of the 12 carpenters choose to wear a device on 4 or more days out of 5, self-selected, with written instruction to supervisors that wear is not directed and not tracked for discipline; (b) zero device-attributable hazard events (snag, restricted egress, trip, dropped tool); (c) median end-of-shift self-reported shoulder and neck discomfort on a 0–10 scale, collected daily, falls at least 2 points against a two-week pre-device baseline; (d) measured task duration on overhead backing no worse than baseline |
| Cost cap | $58,000 all in — devices, fitting, training, Dani's hours. Stop at the cap regardless of results |
| Kill criterion | Any device-attributable near-miss, or voluntary wear below 40% at the end of week 4 |
Why it is built this way. Criterion (a) is voluntary adoption, the honest proxy in a short window: a device workers choose to wear is doing something for them, and one they abandon is not. Criterion (c) is subjective and unblinded — say so in the write-up, because a pilot honest about the weakness of its own measure is worth ten that are not. Criterion (b) and the kill criterion exist because the most likely way this technology hurts you is by creating a new hazard while solving an old one.
C3. The drone earthwork volume — week 9 ⭐⭐⭐
Northgate mass excavation: 44,000 CY cut, 12,000 CY fill, 32,000 CY net export. The earthwork subcontract unit price for excavate-load-haul-and-dispose is $9.85/CY.
At week 5 three measurements of cut-to-date were taken: the sub's ticket-based claim of 31,200 BCY; a $450 drone photogrammetric surface at 29,430 BCY; and a $3,100 conventional GPS rover survey at 29,910 BCY. Kestrel paid on the survey.
It is now week 9. The subcontractor claims 40,600 BCY. Lorena Vasquez flies again — same aircraft, same control points, same altitude, same operator, $450 — and the surface returns 39,910 BCY.
(a) Compute the claim-to-drone gap in cubic yards, as a percentage of the drone number, and in dollars. (b) Compute the ratio of this week's gap to the drone-versus-survey spread you observed at week 5, and the same ratio for week 5. What does the comparison tell you? (c) Do you order the $3,100 survey? Show the arithmetic, including your best estimate of what the survey would return if the week-5 bias repeats. (d) Your civil superintendent mentions the pre-construction existing-ground surface was flown before 4,100 CY of topsoil was stripped, and the stockpile currently sits inside the boundary you are measuring. Which direction does that bias the cut, by roughly how much, and what is the correct fix? Assume 25% swell from bank to loose. (e) Write the file note describing this week's measurement, honestly.
Answer
(a) 40,600 − 39,910 = 690 CY = 690 ÷ 39,910 = 1.73%. At $9.85/CY: 690 × $9.85 = $6,796.50.
(b) The week-5 drone-to-survey spread was 29,910 − 29,430 = 480 CY — your empirically observed uncertainty between the two methods, on this site, with this control, in this material.
Week 9: 690 ÷ 480 = 1.44 Week 5: (31,200 − 29,430) ÷ 480 = 1,770 ÷ 480 = 3.69
At week 5 the disagreement was nearly four times the noise. This week it is barely larger than the noise. You cannot distinguish this week's claim from measurement error, and that is the finding.
(c) Do not order the survey. If the week-5 bias repeats — the survey came in 480 CY above the drone — the expected survey result is about 39,910 + 480 = 40,390 BCY. The defensible difference against the claim would then be 40,600 − 40,390 = 210 CY = 210 × $9.85 = $2,068.50.
Expected recovery $2,069 against a $3,100 survey = a net loss of about $1,031, before anybody's time.
Pay the claim, note the measurement in the file, keep flying. This is the drone doing its actual job: it did not win you money this week; it stopped you spending $3,100 to chase $2,069, with a number you can show anyone. Compare week 5, where the same $450 flight justified the survey and identified a $12,707 overclaim (1,290 CY × $9.85). Same technology, same cost, opposite decision, both correct. A screen, not a scale.
(d) The volume is a difference of surfaces: Σ (existing elevation − current elevation) × cell area. Where the stockpile sits, the current surface is above the existing surface, so those cells contribute a negative term. The stockpile suppresses computed cut; your number is understated. On magnitude: the surface method measures the pile's geometric volume as it sits — loose measure. 4,100 CY loose ÷ 1.25 = 3,280 BCY. So the suppression is roughly 4,100 CY of surface volume, corresponding to about 3,280 CY of the bank quantity you pay against.
The fix is not a correction factor applied to the whole site. Exclude the stockpile footprint from the measured boundary and account for the pile separately, or relocate it before the next flight. A blanket adjustment spreads a local error across the whole surface and makes every future number quietly wrong. This is §39.3.2's central warning: half of every volume calculation is the surface you are comparing to, and that error will not look like an error — it will look like a confident number with a decimal point on it.
(e) A model file note:
Week 9 cut-to-date measured by drone photogrammetric surface, flown Tuesday morning of week 9 by Dani Okonkwo, certified operator, using the six ground control points established at mobilization, differenced against the pre-construction existing surface. Result 39,910 BCY against a subcontractor claim of 40,600 BCY — a difference of 690 CY, or 1.7%, within the drone-to-conventional-survey spread of 480 CY observed at week 5. No conventional survey ordered this period; expected recovery does not justify the survey cost. Note for the record: the pre-construction existing surface predates topsoil stripping and the current stockpile lies inside the measured boundary, understating computed cut by approximately 4,100 CY of surface volume; the boundary is to be revised before the week 13 flight. Payment quantities remain governed by the measurement method named in the subcontract.
That last clause is the one that matters. See Chapter 38: decide how a quantity gets measured before there is money riding on the answer.
C4. The maturity-sensor float test at $5,150/CD ⭐⭐⭐
Jamal Foster wants to extend Kestrel's concrete maturity program beyond the foundation walls and grade beams. Four candidate scopes on Northgate. Extended general conditions run $5,150 per calendar day; burdened craft rate $54/MH; crews work 10-hour days. The program — loggers, mix-specific calibration for two mixes, setup, and crew training — costs $14,800.
| Scope | CD removed from the scope | Float on the successor path | Crew | Crew-days saved |
|---|---|---|---|---|
| A. Foundation walls and grade beams (14 placements, 2 form sets) | 9 CD | 22 CD | 7 | 7 WD |
| B. Elevator and stair pit walls (6 placements, 1 form set) | 8 CD | 3 CD | 5 | 6 WD |
| C. Level 2–4 lightweight topping placements | 5 CD | 0 CD — on the critical path | 6 | 4 WD |
| D. Exterior site retaining walls (8 placements, 2 form sets) | 6 CD | 31 CD | 5 | 5 WD |
(a) For each scope, how many calendar days actually move substantial completion? (b) Value that movement at the extended general-conditions rate. (c) Value the labor saving in each scope. (d) Total return, and net after program cost. (e) Which scope has the largest saving on paper and the smallest real one? Which has the smallest on paper and among the largest real? (f) Had you built the business case on the sum of the paper days, by how much would you have overstated the schedule value — in days and dollars? (g) Northgate's total daily exposure is $10,650/CD ($5,150 extended GC plus $5,500 liquidated damages). Why is $5,150 the correct rate here and not $10,650?
Answer
(a) The float test. Days saved inside available float restore float; only days beyond it move the finish.
| Scope | CD saved | Float | CD that actually move the finish |
|---|---|---|---|
| A | 9 | 22 | 0 |
| B | 8 | 3 | 8 − 3 = 5 |
| C | 5 | 0 | 5 |
| D | 6 | 31 | 0 |
| Total | 10 CD |
(b) B: 5 × $5,150 = $25,750. C: 5 × $5,150 = $25,750. A and D: $0. Total schedule value = $51,500.
(c) Man-hours saved = crew-days × crew × 10 hours.
| Scope | Arithmetic | MH | Value at $54/MH |
|---|---|---|---|
| A | 7 × 7 × 10 | 490 | $26,460 |
| B | 6 × 5 × 10 | 300 | $16,200 |
| C | 4 × 6 × 10 | 240 | $12,960 |
| D | 5 × 5 × 10 | 250 | $13,500 |
| Total | 1,280 | $69,120 |
(d) $51,500 + $69,120 = $120,620, less $14,800 = net $105,820. Note where it came from: the majority is labor, not schedule. The labor saving is real on every scope regardless of float, because you performed the same work with fewer crew-days wherever the days landed. Float governs only the schedule half.
(e) Scope A — largest on paper (9 CD), smallest real (0), sitting behind 22 days of float. That is exactly §39.7.2's Northgate finding: foundations complete July 3, Year 1; steel not scheduled until August 4. Scope C — smallest on paper (5 CD) and worth every day, because it is on the critical path.
(f) Paper days = 9 + 8 + 5 + 6 = 28 CD = $144,200. True value $51,500.
Overstatement = 18 CD = $92,700, a claim 2.8 times the truth.
That is the business case that makes an operations executive stop believing technology proposals, and one question prevents it: is this scope on the critical path?
(g) Because liquidated damages are only owed if you finish late. On a job projected to finish on time there are no LDs to avoid — only extended general-conditions burn you stop paying. Applying $10,650 claims credit for avoiding a penalty that was never going to be assessed. The full rate is right when you are valuing days on a job already projected late, which is how the steel-delay acceleration analysis was framed. Same project, same two rates; the schedule decides which applies.
C5. The reality-capture business case — break-even, not ROI ⭐⭐
Grace's expansion proposal — a second scanner, two software seats, training, a half-time technician — costs $148,000 in year one and $61,000 per year thereafter. Kestrel's evidence: 11 scanned scopes last year, 6 documented catches, aggregate avoided rework estimated at $214,000.
(a) Three-year cost of the expanded program. (b) The annual avoided-rework figure at which it breaks even over three years. (c) Add Proposal A from C1 — as-built sold at $66,000 with $44,000 of delivery cost, three sales over three years. Recompute break-even. (d) Why is break-even a more defensible number to put in front of Owen Baptiste than a return-on-investment figure computed from the $214,000?
Answer
(a) $148,000 + $61,000 + $61,000 = $270,000.
(b) $270,000 ÷ 3 = $90,000 per year — well under half the $214,000 already demonstrated.
(c) Margin per sale = $66,000 − $44,000 = $22,000; three sales = $66,000. Adjusted three-year net cost = $270,000 − $66,000 = $204,000; break-even = $68,000 per year.
(d) Because "avoided rework" is a counterfactual, and counterfactuals run optimistic. The $214,000 estimates what six catches would have cost had nobody caught them, and every one of those estimates was made by someone who wanted the program to look good — not dishonestly, but people estimate hypothetical disasters generously. A CFO is right to discount it.
Break-even needs none of that. It says: here is the cost; here is the avoided rework required for it to be free; you decide whether that is plausible. You have moved the uncertain estimate off your slide and onto the reader's judgment — more honest, and far more persuasive.
And notice what Proposal A really accomplishes. The $66,000 matters less than the fact that Kestrel stops giving away a deliverable owners value — the misalignment §39.9.3 names explicitly. Pricing it is the fix.
Part D — Judgment and Ethics ⭐⭐⭐
D1. The worker-monitoring question. Bea Salgado wants heat-strain monitoring on summer exterior work. The device she likes also logs continuous location, because that is how it works — you cannot buy the heat function alone. Roughly 80% of the people who would wear it are subcontractor employees.
Answer all five in writing, specifically enough that somebody could act on them:
- Write the written commitment you would hand every wearer before the first device goes on a belt: exactly what is collected, exactly who can see it, exactly what it will and will not be used for, and how long it is retained. Include the commitment §39.7.4 says earns you credibility.
- Write the plain-English subcontract language placing this on a subcontractor's employees. Is a subcontractor's consent the same thing as their employee's consent? Note where this depends on jurisdiction and on any applicable collective bargaining agreement, and say who you would ask.
- A worker declines to wear one. What happens? Give your actual answer, not the comfortable one, then say what it implies about whether the program is voluntary.
- Six months in, a subcontractor's employee is terminated for a reason unrelated to safety and their employer asks Kestrel for that worker's location history to support the decision. You committed in writing that the data would not be used in discipline. What do you do, and what does it cost you either way?
- Nadia asks whether it would be simpler to require it in the subcontract and move on. Give her the argument against management by fiat. Then: if the vendor will not sell a version without location logging, does that change your recommendation?
D2. Computer vision and the safety count. A tool that reviews site photographs and flags missing personal protective equipment is proposed on your job. It works, and it produces a weekly violation count by subcontractor.
Using Chapter 24's threshold concept — safety is a property of the production system, not a rulebook — argue whether this makes your site safer. Then design the version you would deploy: what output must it produce, who receives it, and what decision does it feed? Finally, describe what happens to your near-miss reporting rate once the violation count becomes an enforcement instrument, and explain the mechanism.
D3. The honest schedule claim. The Northgate corridor program compressed overhead rough-in 19 work days, which produced 9 calendar days at substantial completion. Marketing wants "19 days faster" in a proposal. Operations wants "9 days." A competitor's proposal on the same pursuit claims a percentage improvement with no basis stated.
Where is the line between optimistic and misleading? Write the sentence you would actually put in the proposal and defend it. Then the harder question: what do you do when the honest number loses you the job?
D4. Killing a pilot without killing the appetite. Bea Salgado championed the wearable pilot. It was well run, it answered its question, and the answer was no. Nadia killed it in about eleven seconds in a room where Bea was sitting.
Write what you would say in that room, in Nadia's position, in under sixty words. Then: what specifically must not happen to Bea in the next twelve months, and what happens to every future pilot report at Kestrel if it does? Describe the failure in terms of the data you would subsequently receive, not in terms of morale.
D5. What the as-built model warrants. Kestrel hands Meridian a scanned as-built model at closeout. Three years later a facilities engineer designs a small renovation off it, relies on a duct location the scanner could not actually see — it was behind a fire-rated shaft wall closed before the scan — and the contractor hits it.
Who is exposed, and to what? Write the reliance-and-accuracy statement that should have accompanied the model, in language a facilities engineer would actually read. Then the commercial question: does that statement make the deliverable worth less to the owner, and if so, is that a reason to leave it off?
Part M — Mixed and Interleaved Practice ⭐⭐–⭐⭐⭐
M1. With Chapter 35 — coordination LOD versus fabrication LOD. Your BIM execution plan specifies a level of development sufficient for clash detection in all overhead areas. Your mechanical subcontractor now wants to prefabricate two corridors from it. List everything that must change in the plan, name who pays for each change, and identify the single most likely place a dimension will be wrong. Then explain, using the level-of-development rule, why a model you cannot trust is worse than no model at all in a fabrication context specifically — what does everybody do differently when there is no model?
M2. With Chapter 25 and Chapter 26 — the record as evidence. A subcontractor claims another trade damaged its conduit above a corridor ceiling and submits a change order request. You have 360-degree captures of that corridor on four consecutive weeks. Describe how you would use the record, then write the substantive paragraph of your response — the one that ends the conversation without starting a fight. What would you do differently if the captures had been monthly? And what does the existence of that record change about how you write, given that everything you write may be read in a dispute?
M3. With Chapter 3 — delivery method as a technology gate. Rank design-bid-build, CM at Risk with a GMP, and design-build by how available multi-trade prefabrication is under each, and explain the ranking in terms of mechanism — what each method specifically gives you or denies you. Then apply it: the Willow Street Community Center is design-bid-build with a municipal owner. Name the single structural reason a multi-trade rack program is unavailable there, and say whether any amount of technical skill compensates for it.
M4. With Chapter 14 and Chapter 29 — float as the exchange rate. A vendor's business case values every day their product saves at your full daily exposure rate. Write the four questions you would ask to test that claim, in order. Then explain, to a reader who has never taken a scheduling course, why a day saved on an activity with 22 days of float is worth zero dollars of extended general conditions — and why it is nonetheless not worth nothing, using what happened at Northgate five weeks after the foundations finished.
M5. With Chapter 20 and Chapter 21 — measurement versus management. Chapter 20's position is that productivity is a management variable more than a worker variable: sequence, access, staging, crew size, overtime, stacking, supervision ratio, and rework are all decided by the project team. Sort every technology in Chapter 39 into two lists — those attacking a management variable and those attacking measurement. Why does the second list hold most of this chapter's failures, and what does that tell you about your next technology dollar? Use equipment telematics as your test case and say which list it belongs on, because the answer is not obvious.
M6. With Chapter 38 — production versus assembly. Explain why machine control succeeded first and hardest in heavy civil rather than in building construction, in terms of production thinking versus assembly thinking. Then explain the sentence the chapter says is worth more than any other in it: prefabrication converts assembly work into production work. Use the Northgate corridor numbers — 59 modules, a learning curve, 4,300 shop man-hours — to show the mechanism concretely. Finally, say why the rebar-tying robot that failed on a building slab might succeed on the Cottonwood Creek Bridge Replacement, and what you would check before believing it.
Part E — Research and Extension ⭐⭐⭐⭐
E1. Find out what it actually takes to fly. Chapter 39 gives no rule numbers, because commercial unmanned aircraft requirements vary by country and change. Find the current answer for your jurisdiction, by function, and write a one-page memo covering: operator certification and how long it takes to obtain; aircraft registration and marking thresholds; airspace authorization, including the process near a hospital helipad; operational limits on flight over people, beyond visual line of sight, at night, and by altitude, and which have waiver paths; and the insurance question — call a broker and ask whether your general liability policy excludes aircraft operations. Cite only the aviation authority's own current material, and put the date you checked at the top. That last habit is the most valuable thing in this exercise.
E2. Read the productivity argument yourself. §39.1 asserts that construction labor productivity has lagged manufacturing for decades, that the direction is not seriously contested, and that the size genuinely is. Go to the Bureau of Labor Statistics industry productivity program and find the construction series. Then find one study arguing the gap is overstated and one arguing it is real and large. Write 400 words on why the measurement is hard — what does "output" mean for a product that is different every time, built outdoors, by a temporary coalition, with quality and complexity changing across decades? Do not resolve the argument; describe it accurately enough that a reader could hold both positions.
E3. Follow the capital. §39.10 argues the most credible adoption signal is not a survey but a subcontractor's own money. Find a mechanical, electrical, or sprinkler contractor in your market and ask three questions: do you have a fabrication shop, what did you build it for, and roughly what percentage of your work goes through it? Then ask the fourth, which is the real one: what does a general contractor have to do for you to make prefabrication work? Write the answer down verbatim. It will be about design freeze, model quality, or getting paid for inventory — and whichever they name is the thing to fix first on your next project.