Capstone: the Northbridge bridge file

This final lesson pulls the course together by following the Clearwater pedestrian bridge from the first phone call to the end of the corrosion dispute, stopping at each point where law, ethics or professional practice shaped what Northbridge did. Along the way it covers the one major topic not yet treated in depth: how construction and professional disputes are resolved. The goal is to see the subjects of the earlier lessons not as separate chapters but as one continuous set of responsibilities that an engineer carries through a project.

Stage 1: winning and signing the work

The project begins with a request for proposals from the Rural Municipality of Clearwater. Northbridge's first questions are professional ones (lesson 1): does the firm hold a certificate of authorization in Saskatchewan, and does it have engineers licensed there and competent in steel bridge design? Maya Okafor is, so the firm proposes. The municipality uses qualifications-based selection with a fee envelope (lesson 5), and Northbridge is ranked first.

The municipality's draft agreement contains a highest-standard clause, an unlimited indemnity with a duty to defend, and unlimited liability. Northbridge negotiates a reasonable standard of care, an indemnity limited to its own negligence, and a cap equal to its insurance limit, applying to claims in contract, tort or otherwise (lessons 3, 6 and 7). It confirms in writing that the client declined full-time field review in favour of periodic visits, and its proposal and the signed agreement are reconciled into one document before work starts, avoiding the counter-offer trap of lesson 3.

Stage 2: design and the duty to say no

During design, three ethical and practical pressures arrive. The client asks Northbridge to cut the geotechnical investigation to two boreholes to save money; Maya explains in writing why four are needed and what risk the client takes with fewer, and the client agrees to four (lessons 7 and 10). A steel supplier offers Maya's team a weekend at a fishing lodge; the firm's gift policy and the codes require her to decline (lesson 10). And Daniel Ruiz, now an engineer-in-training with growing responsibility, designs the connections under Maya's direct supervision, with every calculation checked before Maya authenticates the drawings (lesson 11).

Daniel's faster bolted connection raises the intellectual property questions of lesson 9. Under his employment agreement the invention belongs to Northbridge, and under its services agreement the client has a licence to use it on this bridge. Because Daniel described it at a conference before anyone filed, the firm has a year to decide whether a domestic patent is worth pursuing, and has already lost rights in absolute-novelty countries.

Stage 3: tender and construction

The municipality tenders construction (lesson 5). Northbridge recommends disqualifying the two lowest bids as non-compliant and awarding to Riverbend, the lowest compliant bidder, and writes down its reasons. The construction contract is CCDC 2 with supplementary conditions Northbridge drafted and explained to the client (lesson 6).

During construction Northbridge wears two hats: designer, and the owner's contract administrator who must decide fairly between owner and contractor (lessons 4 and 6). It grants the contractor time for a flood and an owner design change, assesses liquidated damages correctly for the remaining delay, and reviews shop drawings for general conformance only, while the fabricator's engineer takes responsibility for the connection details it designs. Daniel's warning about the temporary crossing (lesson 8) and Maya's decision to hold a pour until cracking near two connections was understood (lesson 10) are the two moments where safety depended on an engineer refusing to let schedule decide.

Stage 4: the claim

Five years after opening, corrosion closes the bridge and the municipality sues Northbridge, Riverbend and the steel fabricator for about $900,000 (lesson 7). Northbridge notified its insurer when the inspection report first arrived, as its claims-made policy requires. Its records, kept since the project, contain the drainage calculations, the shop drawing review comments and the field reports. The question now is how to resolve the claim.

How disputes are resolved

Disputes can be resolved in several ways, usually arranged as an escalation ladder in the contract. Each step up costs more, takes longer and hands more control to a third party.

• Negotiation between the parties, often at escalating levels of management, is the cheapest and fastest. Most standard forms require it first.
• Mediation brings in a neutral third party who helps the parties reach their own settlement but cannot impose one. It is confidential, flexible and usually concluded in a day or two of meetings after weeks of preparation. Mediation succeeds in a large majority of construction cases.
• Adjudication is a fast, interim decision by an adjudicator, usually within weeks, binding unless and until the dispute is finally decided by arbitration or a court. Ontario introduced statutory adjudication for payment disputes in its Construction Act in 2019, Saskatchewan's prompt payment and adjudication amendments to The Builders' Lien Act took effect in 2022, and the federal Prompt Payment for Construction Work Act applies to federal construction projects. Dispute boards, standing panels of neutral experts that follow a project and decide issues as they arise, serve a similar purpose on large projects.
• Arbitration is private adjudication by an arbitrator or panel the parties choose, under provincial arbitration statutes such as Saskatchewan's Arbitration Act, 1992, or the CCDC 40 rules for construction arbitration. The award is final and binding, with very limited rights of appeal. Arbitration is confidential, lets parties choose technically expert arbitrators, and can be faster than court, though large arbitrations can be just as expensive.
• Litigation in the courts is public, follows formal rules of procedure and evidence, allows appeals, and can join every party to a multi-party dispute, which arbitration often cannot do without everyone's agreement. It is also slow: complex construction cases can take years to reach trial.

The engineer's role in disputes takes several forms. As a party, the engineer must preserve records and cooperate with its insurer. As contract administrator, it makes the first-level decisions under the contract and must do so impartially. And as an expert witness, an engineer gives opinion evidence to help a tribunal understand technical issues. An expert's duty is to the tribunal, not to the party paying the fee: the opinion must be independent, objective and within the expert's competence, and courts in both countries reject experts who act as advocates. Codes of ethics apply with full force to expert testimony.

Choosing a path for the corrosion claim

Northbridge's insurer, its defence lawyer and Maya sit down to compare the options. The program uses their estimates: the probability that Northbridge would be found negligent, its likely share, the worst case, and the cost and time of each route.

# Northbridge's insurer compares ways to resolve the corrosion claim.
# All figures are estimates in dollars; probabilities are judgment calls.
SHARE = 243_000        # Northbridge's likely share if found liable
WORST = 607_500        # worst-case payment (joint and several: the contractor may be unable to pay)
P_LIABLE = 0.60        # chance Northbridge is found negligent
P_WORST = 0.15         # chance of the worst case, given liability

verdict = P_LIABLE * ((1 - P_WORST) * SHARE + P_WORST * WORST)
options = {
    # name: (expected payment, own legal and expert costs, months)
    "Pay the demand": (180_000, 15_000, 2),
    "Mediate":        (150_000, 40_000, 8),
    "Arbitrate":      (verdict, 150_000, 18),
    "Go to trial":    (verdict, 260_000, 40),
}
print(f"Expected award if decided: {verdict:,.0f}")
for name, (pay, cost, months) in options.items():
    print(f"{name:14} payment {pay:9,.0f}  costs {cost:8,.0f}  total {pay + cost:9,.0f}  ~{months} months")
best = min(options, key=lambda n: sum(options[n][:2]))
print(f"Lowest expected total: {best}")

Expected award if decided: 178,605
Pay the demand payment   180,000  costs   15,000  total   195,000  ~2 months
Mediate        payment   150,000  costs   40,000  total   190,000  ~8 months
Arbitrate      payment   178,605  costs  150,000  total   328,605  ~18 months
Go to trial    payment   178,605  costs  260,000  total   438,605  ~40 months
Lowest expected total: Mediate

Mediation comes out lowest, slightly ahead of simply paying the municipality's demand and far ahead of a full arbitration or trial, whose legal and expert costs swamp any difference in the award. The numbers are judgment calls, but they explain why so few construction claims go to trial. The joint and several worst case matters too: if Riverbend's insurer is thin, Northbridge could end up paying far more than its share, a risk that settling removes.

The mediation succeeds. Northbridge's insurer pays its share, the contractor and fabricator pay theirs, the municipality accepts responsibility for its missed inspections, and nobody admits liability. Northbridge also does something no settlement requires: it reviews its standard drainage details, issues a technical bulletin to its engineers, and adds a corrosion detailing check to its design review checklist. That is the professional response to a failure, and it is what regulators and courts look for when they ask whether a firm learns from its mistakes.

What runs through the whole file

Looking back across the file, a few habits did most of the work.

• Know the rules of the place: licences, firm authorization, procurement law, safety and environmental statutes, limitation periods and liability rules all depend on the province or state.
• Get it in writing: proposals reconciled with agreements, scope changes authorized, client decisions against advice confirmed, warnings documented.
• Match contracts to insurance: standard of care, indemnities and caps that the firm can actually insure.
• Check, and check changes: independent checking before sealing, and fresh calculations whenever a design or a shop drawing changes an assumption.
• Put the public first, early: decline gifts, hold the pour, stop the shoring removal, recommend the extra boreholes. Small, early decisions avoid large, late failures.
• Keep records and learn: records won the arguments years later, and a firm that improves its practice after a claim protects the next public it serves.