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 discusses incorporating correctness into system design and language choices, and says there is no simple defense against attacks involving compromised software packages.
The Pragmatic Engineer has published an interview with Sean Allen, a distinguished engineer at Antithesis, about software correctness and the challenges of preventing software supply chain attacks. Allen discusses incorporating correctness into system design and programming languages, rather than relying only on testing at the end of development.
Allen defines correctness in practical terms: software should do what its users and operators expect, expressed where possible through invariants—guarantees about how a program behaves. He also cautions that teams may not know those expectations clearly before a system is built. In his words, software is “correct enough” when it serves its business purpose and does not wake someone at 2 a.m. with an outage.
The interview challenges reliance on manual testing as the main route to reliability. Allen favors preventing invalid states from being represented in the first place, including through language design and strong type systems. He also discusses repeated execution of defined workloads as a way to test behavior; one example is the Bank Test, used to stress database isolation guarantees. These approaches are presented as part of a broader strategy, not as a claim that tests have no role.
On supply chain security, Allen describes a difficult trade-off: languages designed with strong security protections may have limited adoption, while widely used languages often operate on assumptions of trust in their package ecosystems. A compromised package can execute malicious code on a developer’s machine or a production service. Allen’s conclusion, as reported in the interview, is that there is no single or simple fix.
Correctness Before Release Checks
Allen argues that reliability work should include design decisions made during development, rather than relying only on checks at the end of the process. He discusses correctness as part of design, including language constraints and validation that can identify issues while software is being built. These approaches aim to prevent some classes of errors, though the interview does not compare their effectiveness with other methods.
Allen also notes that organizations may rely on code written and distributed by parties outside their direct control. A package compromise can allow malicious code to reach machines or services. His comments do not prescribe a universal defense; they identify ecosystem trust as a security risk without presenting a single solution.
The interviewer says concern about more code being produced, including by AI agents, motivated the conversation. That is the interviewer’s rationale, not evidence in the source that these tools have caused a measured increase in defects.
Products Worth Considering
Samsung Galaxy Tab A9+ 11-Inch Android Tablet, 64GB, Graphite
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
Bose TV Speaker Compact Soundbar with Bluetooth and HDMI-ARC, Black
As an affiliate, we earn on qualifying purchases.
Allen’s Engineering Background
Allen works on software correctness full time at Antithesis. The source describes his experience building large distributed systems, stream processing and confidential computing. He is also a core contributor to Pony, a programming language designed for concurrency. The interview draws on those areas to discuss how language features and system design can affect reliability.
The article ranges beyond supply chain security. Allen contrasts language families, discusses workloads as an alternative or complement to unit-testing practices, and challenges the idea that quality should be enforced only through late-stage CI/CD gates. He also argues that technical debt may not always need to be repaid: during a rewrite, a team may leave behind old compromises rather than reproduce them. These are Allen’s views in an interview, not findings from a comparative study.
The Pragmatic Engineer notes that Antithesis sponsors its podcast, while stating that the interview itself was not sponsored. The publication also announced a separate return of The Pragmatic Summit, scheduled for 10 February in San Francisco, with tickets priced at $1,000. That event announcement is distinct from the interview’s technical discussion.
““It’s down to whether my program does what I expect.””
— Sean Allen, as quoted in The Pragmatic Engineer
Limits of the Security Advice
The supplied interview excerpt does not set out a detailed defense plan, quantify the frequency or cost of package compromises, or compare the effectiveness of specific security measures. Allen’s assessment that no simple fix exists should not be read as a claim that organizations have no ways to reduce risk. The excerpt also does not provide test results establishing that one programming language or testing method is superior across all systems.
The source material includes only the opening of the interview and a summary of topics; it does not include the full discussion or additional detail on the language comparisons, garbage collection or the specific security properties of the E language. The publication date is also not specified in the material provided.
Practical Questions for Engineering Teams
The interview does not announce a product, policy change or follow-up milestone. It raises questions for teams to consider: which behaviors their systems must guarantee, whether invalid states can be ruled out earlier, and how repeated workloads can test important system properties. Teams may also review their reliance on third-party packages and the trust assumptions those dependencies introduce.
Those steps are applications of the discussion, not actions Allen says every organization has already taken. The source provides no announced schedule for further reporting on the subject. The Pragmatic Engineer separately says more speakers are expected for its 10 February Summit in San Francisco, but that announcement is not presented as a continuation of Allen’s interview.
Key Questions
What is the news?
The Pragmatic Engineer published an interview with Sean Allen about software correctness, testing and software supply chain security.
How does Allen define software correctness?
Allen describes it as software doing what is expected of it. He discusses expressing those expectations through invariants, or guarantees about program behavior.
What does Allen say about supply chain security?
He says it is difficult to defend against attacks involving hijacked packages and that there is no single or simple fix. The supplied material does not give a detailed defense checklist.
Does the interview argue that testing is useless?
No. Allen is described as skeptical of manual testing as the main way to ensure correctness. He favors preventing invalid states through design and using defined workloads to exercise systems repeatedly.
Source: newsletter.pragmaticengineer.com
