TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
The Pragmatic Engineer published an interview with Sean Allen, a distinguished engineer at Antithesis, about software correctness and supply chain security. Allen argues for designing systems around invariants and repeatable workloads, and says there is no single, simple defense against attacks on software dependencies.
The Pragmatic Engineer has published an interview with Sean Allen, a distinguished engineer at Antithesis, about software correctness and supply chain security. Allen argues that correctness starts with defining what a system must guarantee, while defending against compromised software dependencies remains difficult because popular programming languages often assume a trusted ecosystem.
Allen describes correctness through invariants: guarantees that software behaves as expected. In the interview, he says teams may not know the full definition of success before a system is built, even if they can recognize failures. His practical standard is software that serves its business purpose without causing an overnight outage.
He is skeptical of relying on manual testing as the primary way to establish correctness. The interview presents his preference for making invalid states difficult or impossible to express in a programming language, then checking behavior through repeatable workloads. One example discussed is the Bank Test, a workload used to stress database isolation guarantees. Allen also favors feedback on quality throughout development over treating quality as a gate added late in a CI/CD process.
On supply chain security, Allen describes a risk in which an attacker takes over a software package and malicious code runs on a developer’s machine or a service. He says languages designed with strong security protections, such as E, have not become popular, while widely used languages often presume that the surrounding package ecosystem is trustworthy. The interview does not identify a single defensive measure that resolves the problem.
Correctness Begins Before Testing
The discussion matters to engineering teams because software failures can affect service availability, data integrity and the time developers spend responding to incidents. Allen’s emphasis on invariants and language design puts part of the correctness problem before test execution: teams need to specify expected behavior and choose tools that help prevent invalid conditions.
His comments on supply chain risk point to a different challenge. Software increasingly relies on third-party packages, but the interview frames defenses as constrained by the design assumptions of popular languages and ecosystems. Allen’s assessment is not a security audit or a new technical finding; it is an experienced engineer’s view that teams should not expect a simple fix. Readers should treat it as a warning about the limits of relying on ecosystem trust alone, not as evidence that a particular package or language is compromised.
The article also challenges the idea that quality can be handled through a final approval step. Continuous feedback and workloads that exercise system behavior offer a different approach, though the source does not provide comparative data showing that one method produces fewer defects than another.
Products Worth Considering
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
DJI Osmo Action 4 Essential Combo Waterproof 4K Action Camera
Amazon Kindle Scribe (3rd Generation), 64GB, 11-inch Display, Graphite, with Premium Pen
As an affiliate, we earn on qualifying purchases.
Allen’s Work on Correctness
Allen works on software correctness full time at Antithesis. The interviewer, Gergely Orosz of The Pragmatic Engineer, says he sought Allen’s perspective because of his experience with confidential computing at Microsoft, distributed systems and the Pony programming language, where Allen is a core contributor. Pony is designed for concurrency.
The conversation ranges beyond security. Allen discusses strong type systems, technical debt, garbage collection and the value of spending more time on design and team alignment before implementation. The article summarizes his view that technical debt does not always need to be repaid, particularly when a system is rewritten after becoming popular. These points are presented as his perspective, not as universal rules for software projects.
The publication says Antithesis sponsors its podcast, but states that the interview itself is not sponsored. The source also announces that The Pragmatic Summit is scheduled for February 10 in San Francisco, with tickets priced at $1,000 and applications open for senior engineers and engineering leaders. That event notice is separate from Allen’s interview.
““It’s down to whether my program does what I expect.””
— Sean Allen, distinguished engineer at Antithesis
Limits of the Security Assessment
The interview gives Allen’s assessment but does not offer a technical comparison of language defenses, package ecosystems or attack rates. It does not name a specific incident, affected dependency, organization or mitigation, and it does not establish that any one language or practice can prevent supply chain attacks.
The source also does not provide controlled results comparing manual testing, type systems, workloads or late-stage quality gates. The Bank Test is described as an example of a workload, but no test results or implementation details are included in the supplied material. The publication date is not stated in the source excerpt, and it is unclear whether Allen’s comments reflect any developments after the interview was prepared.
Applying the Ideas in Practice
The source does not announce a product, security policy or follow-up research arising from the interview. The immediate next step for readers is to consult the full conversation for Allen’s explanations of language design, testing and runtime performance; the publication also says the discussion is available on YouTube.
For engineering teams, the questions raised are practical: which behaviors must always hold, how can workloads exercise those guarantees repeatedly, and what trust assumptions are embedded in their languages and dependencies? The interview does not prescribe a common checklist or timeline, so any changes would depend on each system and organization.
Key Questions
What is the news development?
The Pragmatic Engineer published an interview with Sean Allen about software correctness, testing and supply chain security.
How does Sean Allen define software correctness?
Allen describes it as software doing what is expected, expressed through invariants or guarantees. He also gives a practical standard: the system should serve its business purpose without causing an outage.
Why does Allen question manual testing?
He argues that teams should prevent invalid states from being expressible where possible and use repeatable workloads to examine system behavior, rather than treating manual testing as the main proof of correctness.
What does Allen say about supply chain security?
He says attacks involving compromised software packages are difficult to defend against, in part because popular languages often assume a trustworthy ecosystem. He identifies no single or simple fix.
Was the interview sponsored by Antithesis?
The Pragmatic Engineer says Antithesis sponsors its podcast, but states that this interview was not sponsored.
Source: newsletter.pragmaticengineer.com
