Written in a personal capacity; views expressed are my own. For transparency, I am Chief Scientific Adviser at the Department for Digital, Culture, Media and Sport (DCMS), which now has responsibility for the Women in Tech Taskforce discussed below.
One of the most recognisable photographs in the history of computing shows Margaret Hamilton standing beside a towering stack of printed listings of the software developed for NASA’s Apollo programme. Taken at MIT in 1969, it has become an enduring image of a pioneering woman in computer science, although it is worth remembering that the software represented the work of two teams she led, involving more than 400 people. Hamilton’s achievements were extraordinary, but the photograph captures only part of what made her work so consequential.
Hamilton, who died on 30 September 2026 at the age of 90, helped establish software engineering as a serious technical discipline at a time when software itself was often poorly understood and its practitioners struggled for recognition. Having joined MIT’s Apollo project in 1965 as its first programmer, she went on to lead the development of onboard flight software for the Command and Service Module and Lunar Module, working on some of the most demanding computational problems of the space age. Her contributions encompassed software architecture, error prevention, fault tolerance and the management of complex engineering systems, with methods and principles that remain highly relevant more than half a century later.
The best-known example came during the Apollo 11 lunar landing in July 1969, when the guidance computer began reporting alarms shortly before touchdown. A problem involving the rendezvous radar and its associated switching arrangements overloaded the processor, threatening its ability to manage the landing, but the computer’s priority-driven executive, originally devised by Hal Laning and incorporated into the flight software developed under Hamilton’s leadership, allowed it to recover and continue executing mission-critical tasks. The flight controllers could proceed because the system had been engineered to manage an overload rather than fail indiscriminately under pressure. The NASA Apollo 11 Lunar Surface Journal’s account of the program alarms provides a particularly revealing technical explanation of the priority-driven executive, restart mechanisms and the decisions involved in developing reliable software within the severe computational constraints of the Apollo Guidance Computer. It remains a remarkable example of designing for resilience, anticipating failure and understanding how software must behave within a wider operational system.
Another episode illustrates her engineering judgement just as clearly. During testing for Apollo, Hamilton’s young daughter Lauren inadvertently activated a pre-launch program while using a command module simulator, disrupting its navigation system. Hamilton recommended a software safeguard to prevent the same mistake occurring during a mission, but the proposal was initially rejected because the astronauts were considered too well trained to make such an error. When Jim Lovell inadvertently triggered the same problem during Apollo 8, her team had to help recover the situation, and the proposed changes were subsequently incorporated. The incident, recounted in MIT’s retrospective on Hamilton’s career, exposed the danger of designing around assumptions about how people ought to behave, however experienced or highly trained they might be, rather than accounting for what can happen in practice.
Hamilton was equally determined that the intellectual and engineering work involved should receive proper recognition. Reflecting on the origins of the term software engineering in a 2018 interview with El País, she recalled that her insistence on treating software as an engineering discipline had initially been regarded as amusing, even radical. She wanted software, and the people developing it, to receive the same professional respect as other forms of engineering, while recognising its integral place within the wider systems engineering process. She subsequently carried those principles into her work on software reliability, systems modelling and the companies she founded after Apollo.
That insistence looks particularly prescient today. Software underpins critical infrastructure, healthcare, transport, financial services, communications and digital public services, while AI is changing how increasingly complex systems are developed and deployed. On 21 October, the National Cyber Security Centre issued urgent advice following the disclosure of two critical vulnerabilities in PHP, a programming language underpinning millions of websites. The need for urgent security updates prompted precautionary shutdowns of some government and commercial services while fixes were applied. AI-assisted security research helped establish how the vulnerabilities could be exploited, while the PHP Foundation’s account illustrates the considerable engineering and organisational challenges involved in maintaining mature software, coordinating security disclosures and deploying fixes across widely distributed systems. The incident also exposes the dependence of the global digital economy on the resources and institutional capacity needed to sustain critical open-source infrastructure.
The growing use of AI-assisted programming brings us back to Hamilton’s insistence on engineering discipline. In our recent paper, Responsible Vibe Coding: Architecture, Opportunities, and Research Agenda (Elgendy et al., 2026), we examine the opportunities created by natural-language, intent-driven software development alongside the challenges of reliability, security, governance and meaningful human involvement across the development lifecycle. Generating code may be becoming easier, but developing dependable software still requires rigorous requirements analysis, verification and validation, testing, maintainability and operational responsibility. Those obligations become more important as development accelerates and the relationship between the people specifying a system and the artefacts ultimately deployed becomes more complex.
These are questions of professional competence and public responsibility that institutions such as BCS, the Chartered Institute for IT, have long sought to address through professional standards, ethical conduct and the development of the computing profession. They also connect with the legacy of Tony Hoare and his contributions to formal reasoning, verification and dependable software. Hamilton and Hoare approached different problems, but both helped establish the intellectual and technical foundations on which modern software engineering depends.
Hamilton’s death has rightly prompted renewed recognition of those achievements. It also offers an opportunity to examine the profession she helped establish: who has been able to participate in its development, whose expertise is recognised and trusted, and what responsibilities its institutions and leaders have for the future.
Who gets to become an engineer?
Hamilton became an accomplished engineer and technical leader in a field where women faced considerable barriers to recognition and advancement, and her subsequent visibility has helped challenge assumptions about who belongs in computing. More than fifty years after Apollo, however, the evidence suggests that our progress in widening participation and opportunity has been considerably less impressive than our technological achievements.
The BCS Women in Tech Report 2026 documents the persistence of substantial disparities across the profession. Women accounted for just 22% of the UK’s IT specialist workforce in 2025, a proportion unchanged from the previous year. Representation was especially low in particular engineering occupations, with women accounting for only 7% of IT engineers and 8% of telecommunications engineers over the 2021–25 period. Female IT specialists were also slightly less likely than their male counterparts to occupy managerial or supervisory positions in 2025 (38% compared with 41%), while their median hourly earnings among full-time employees were 10% lower. That last figure is an unadjusted comparison, rather than evidence of unequal pay for equivalent work, but it adds to a picture of substantial differences in participation and career outcomes.
The imbalance extends across engineering more generally. According to EngineeringUK’s latest evidence, women account for only 17% of people working in engineering and technology occupations, using a broader occupational definition than the BCS figures. Perhaps more revealing is the finding from the 2023 Science Education Tracker, published in 2024, that only 16% of girls in England considered engineering a suitable career for someone like them, compared with 44% of boys. The same research found that just 12% of girls felt that being an engineer fitted well with who they were, suggesting that perceptions of professional identity and belonging become established long before recruitment into employment.
The educational evidence reinforces those concerns, although the picture requires careful interpretation. According to the IET’s analysis of the 2026 GCSE results, girls accounted for only 22.8% of UK entries in Computing, the subject label used in the examination statistics for GCSE Computer Science. In England, 36.1% of girls taking the subject achieved grade 7 or above, compared with 28.1% of boys. Across the UK, Computing entries also fell by 7.1%, potentially raising questions about the subject’s appeal and the educational pathways available to young people. Examination results alone cannot explain why pupils make particular choices, and those who enter a subject are not necessarily representative of those who do not, but there is little basis for assuming that the participation gap simply reflects differences in academic capability.
The broader educational pipeline is also changing. In her recent BCS analysis, Computing’s pipeline problem is more nuanced than headlines suggest, Julia Adamson examines declining acceptances to undergraduate computing courses alongside growth in alternative digital qualifications and apprenticeship pathways. Computing degree acceptances fell by 7% in 2026, but the decline was sharper among men than women, while qualifications such as Digital Technology A-levels and digital T Levels have been developing different routes into technical careers. In Wales, Digital Technology A-level entries increased by 26%, even as Computer Science entries declined. These developments are worth examining because a narrow focus on qualifications labelled Computer Science may obscure wider patterns of participation, including routes that lead to specialist technical roles or the application of computing across other disciplines.
Alongside the routes available into technical education, there is a related question about whose contributions are recognised within the curriculum. In an October 2026 study published in the British Educational Research Journal, Cooke et al. (2026) examined science qualification specifications across the UK’s six main examination boards and found that, of 86 scientists referenced at GCSE or equivalent level, only one was a woman: Rosalind Franklin. At A-level or equivalent, just three of 181 scientists referenced were women. Although these findings concern examination specifications rather than everything taught in classrooms, they reveal a striking imbalance in whose contributions are formally recognised, leaving individual teachers to provide a more representative account of scientific achievement. The study concerns science qualifications, but it illustrates a wider issue for computing and engineering: whose achievements young people encounter, and who they might imagine belongs in scientific and technical professions.
Questions of recognition, educational opportunity and career aspirations also run through the research of Professor Louise Archer and colleagues. Drawing on the longitudinal ASPIRES research programme, Archer et al. (2015) further developed the concept of science capital to examine how young people’s science-related knowledge, experiences, social networks and wider resources influence their participation and aspirations. Their research demonstrated that these resources are unevenly distributed, and that young people who enjoy science and recognise its value do not necessarily regard a scientific career as something for them. Educational choices are shaped by more than individual interest or attainment, including whether young people recognise themselves, and are recognised by others, as belonging within a particular discipline. Archer’s subsequent work on justice-oriented science teaching (Archer et al., 2024) has reinforced the importance of recognising and valuing students’ existing experiences and resources, while critically examining the educational cultures and institutional practices that shape participation.
There are important parallels with computing, although the characteristics of the discipline and its educational pathways require specific consideration. In their 2023 paper, Who wants to be a computer scientist?, Hamer et al. examined the aspirations of 4,983 secondary school students in England, drawing on schools with comparatively high uptake of GCSE Computer Science. Their findings highlighted the importance of computing identity, family support and perceptions of the discipline, including substantial differences between girls’ and boys’ aspirations even among pupils who had chosen to study the subject. Read alongside the wider science-capital research, this suggests that improving participation involves examining how computing is presented and experienced, the opportunities young people have to engage with it, and the social and institutional conditions that influence whether they can envisage a future within the profession.
Understanding how schools might change these experiences has itself become an important area of research. The Raspberry Pi Foundation’s Gender Balance in Computing programme, for which Professor Sue Sentance was the strategic research lead, was commissioned by the Department for Education as part of the National Centre for Computing Education. Delivered with partners including the Behavioural Insights Team, Apps for Good and WISE, the programme involved more than 14,500 pupils and 725 teachers in England between 2019 and 2022. It investigated approaches involving collaboration, storytelling, belonging, subject relevance and connections between formal and informal learning, with independent evaluation including a series of randomised controlled trials.
The findings deserve careful attention because they were more complicated than might have been hoped. None of the randomised trials demonstrated a statistically significant impact on their primary outcomes, although several showed positive effects that fell short of statistical significance and the qualitative evidence provided useful insights into classroom experiences. In a subsequent peer-reviewed study of pair programming, Childs and Sentance (2023) likewise reported positive experiences among teachers and pupils without a statistically significant improvement in girls’ attitudes or intentions to study computing further. The disruption caused by the pandemic, the relatively short duration of the interventions and the difficulty of measuring changes in aspirations all matter when interpreting these results, but they underline the importance of evaluating educational initiatives rigorously and being prepared to learn from findings that do not provide straightforward evidence of success.
Research on computer science curriculum reform provides a useful perspective on the institutional conditions that shape participation. In Restart: The Resurgence of Computer Science in UK Schools (Brown et al., 2014), Neil Brown, Sue Sentance, Simon Humphreys and I examined the revival of computer science as a school subject across the UK, including the contribution of Computing At School in developing professional networks and support for teachers. Subsequent research with Faron Moller examined the challenges of translating national curriculum ambitions into classroom practice in Wales. Our more recent comparative study of computing teaching across the UK and Ireland (Whyte et al., 2026) drew on evidence from 475 primary and secondary teachers across five education jurisdictions, identifying differences in specialist provision, resources and assessment pressures. The relationship between institutional practice and girls’ participation is also evident in research by Wong et al. (2026), who examined how 40 English secondary schools presented computing in their curricula and promotional materials. Schools with higher uptake among girls more often emphasised creativity, collaboration and inclusive role models, suggesting that how computing is presented and experienced within schools matters alongside teachers’ capacity to deliver it.
Wales also provides a perspective on the relationship between research, policy development and implementation. In 2013, I co-chaired the Welsh Government’s independent review of ICT and computing in schools with Stuart Arthur and Janet Hayward, which recommended strengthening computer science, developing a cross-curricular approach to digital competence and addressing gender inclusion and perceptions of computing careers. I subsequently led the collaborative development of the Digital Competence Framework and, later, the Science and Technology Area of Learning and Experience for the Curriculum for Wales. These reforms sought to connect technical capability with critical judgement, creativity and an understanding of the social consequences of technology, while recognising the relationships between scientific enquiry, computation, engineering and design. They also illustrate a persistent challenge in curriculum reform: establishing new expectations is only part of the task, and their realisation depends on teachers, schools and the sustained institutional support available to them.
There is a continuing need for rigorous computer science education, alongside broader digital competence and opportunities to apply computing across different contexts. Similar arguments are now visible in England; the November 2025 Curriculum and Assessment Review final report recommended replacing the existing GCSE Computer Science with a broader GCSE in Computing, retaining a strong foundation in computer science while incorporating the critical application and creation of technology, including AI. The Review also identified concerns about gender stereotyping, perceptions of relevance, assessment and the shortage of specialist teachers. The Government has since endorsed the proposed direction of reform, although the new qualification and curriculum changes have yet to be implemented. Together, these developments illustrate a continuing debate about the purposes of computing education and how different learners can be supported to develop the capabilities they need.
This is closely aligned with BCS’s education policy recommendations, which call for qualifications that recognise the needs of future computing professionals, those applying digital technologies in other disciplines, and citizens participating in an increasingly digital society. The challenge is to maintain strong disciplinary foundations while providing multiple credible routes into computing-related study and employment, supported by appropriate teaching expertise, professional development and assessment. That broader conception of educational and professional pathways also has implications for how we recognise technical capability throughout a career, including the role of apprenticeships, professional registration and continuing professional development.
The scale of the challenge has prompted more explicit ambitions for change. In September, EngineeringUK published 40 by 40: Gender equity in engineering and technology education, on behalf of the Gender Pathways collective, proposing that the proportion of girls and young women on engineering and technology education pathways in England should increase from approximately 26% at age 18 to at least 40% by 2040. Its recommendations cover teaching, careers guidance and the cultures of schools and colleges, recognising that sustained improvements require changes across the educational system and clear measures of progress.
The problem continues into higher education and employment; UCL’s Flip the Script research with BCS has documented experiences of marginalisation, stereotyping and harassment among women studying computing, including those who have already demonstrated their capability and commitment to the subject. UCL is working with computing departments in nine UK universities to develop and test approaches to creating more inclusive educational cultures. Comparable questions arise in employment, where recruitment, promotion, access to influential technical work and the recognition of professional expertise can depend heavily on informal networks and established expectations.
The progression from school to university and into professional life is therefore important in its entirety. A student who completes a computing degree but subsequently leaves the sector, an experienced engineer whose technical judgement is repeatedly discounted, or someone returning after caring responsibilities who cannot find a credible route back into senior work all encounter different barriers, requiring different institutional responses. Improving participation in school computing remains essential, but it cannot carry the whole burden of reform.
Professional responsibility and accountability
The computing profession has been shaped by women whose contributions extend across technical research, professional leadership and the development of the institutions that support it. Professor Dame Wendy Hall was a pioneer of hypermedia research, leading the development of the Microcosm system and subsequently helping to establish Web Science as an interdisciplinary field through the founding of the Web Science Research Initiative in 2006, alongside Tim Berners-Lee, Nigel Shadbolt and Daniel Weitzner. Her influence extends well beyond these substantial research contributions: she has served as President of both BCS and ACM, the only person to have held both positions, and has played a leading role in shaping national and international debates on AI, digital governance and the wider responsibilities of computing. Her sustained advocacy for women in the profession has been an integral part of that scientific and institutional leadership.
Professor Sue Black has similarly combined academic expertise with a sustained commitment to developing the computing profession and widening participation. With a research background in software engineering, she founded BCSWomen in 2001, which marks its 25th anniversary this year and has become an established professional network supporting women working in computing. Alongside her influential campaign to save Bletchley Park, she has developed initiatives including #techmums and TechUP, creating opportunities for women to develop technical capabilities and enter or return to computing careers.
Their new BCS podcast, AI&Us, launched earlier this month, is a welcome contribution to public discussions about AI and its implications for society. Hall recalls the observation of her former mentor Karen Spärck Jones that computing was too important to be left to men, extending that argument to the development of AI. It is a useful reminder that questions about whose perspectives influence technological development have been raised by leading computer scientists for decades, even as the technologies and institutions involved have changed substantially.
BCS has also been connecting its work on diversity more explicitly with professional standards and institutional accountability. Its 2025 recommendations with Coding Black Females emphasised fair recruitment and progression, including the use of independent frameworks such as SFIA and Chartered IT Professional registration to support more transparent judgements about skills and competence. Its 2026 Women in Tech report also describes changes to the BCS Code of Conduct and disciplinary procedures, with greater emphasis on inclusive professional behaviour and mechanisms for reporting concerns. Alongside accreditation and professional registration, these are constructive developments that can provide a counterweight to informal or inconsistent decisions about whose expertise is recognised and who is trusted with responsibility, although their practical effect will depend on how they are applied and evaluated.
The Royal Academy of Engineering has been examining related questions about the responsibilities of professional institutions and the cultures they create. Its 2025 Diversity and Inclusion Progression Framework report, developed jointly with the Science Council, presents benchmarking across 45 professional engineering institutions and scientific bodies. The framework considers organisational practice across governance, membership, professional registration, education, accreditation, employment and monitoring, identifying areas of progress alongside continuing weaknesses in some of the activities most directly associated with professional recognition and institutional accountability.
One particularly important finding concerns monitoring and measurement, which remains relatively underdeveloped. The framework relies on organisational self-assessment, making it valuable for understanding institutional maturity and identifying weaknesses, while leaving open the question of whether improved policies and processes are translating into better experiences and outcomes for members. That distinction matters because professional bodies have considerable influence over who is recognised as possessing technical expertise, how professional competence is assessed and which forms of leadership are valued. Decisions about accreditation, registration, professional recognition and fellowship all contribute to shaping a profession’s understanding of excellence, and those processes should be open to the same critical scrutiny that professional institutions encourage within engineering practice.
Government policy is addressing some of these challenges through the Women in Tech Taskforce, established in December 2025 and now co-chaired by DCMS minister Vicky Foxcroft MP and Women in Tech Envoy Dr Anne-Marie Imafidon. Its work encompasses barriers to women’s entry, retention and leadership across the technology sector, including access to training, recruitment, workplace culture and professional progression. The findings from its July 2026 call for evidence, drawing on 570 responses, reinforce many of the concerns identified in the academic and professional literature, although the self-selecting responses should be understood as evidence of experiences and potential barriers rather than representative estimates of their prevalence. The taskforce’s membership also includes Louise Archer, providing a direct connection with the educational research discussed earlier and the importance of grounding policy interventions in a better understanding of participation and opportunity.
The wider programme includes the TechFirst Women’s Programme, supporting women entering, returning to and progressing within technology careers, with an ambition to place at least 300 women into paid roles. The Government’s Digital and Technologies Jobs Plan, published on 21 October, also recognises the importance of broadening participation, supporting teachers and strengthening routes into digital careers. The impact of these measures will need to be assessed beyond initial participation and placement, including whether participants remain and progress in the sector and what can be learned from implementation.
The taskforce’s terms of reference include establishing frameworks for measuring progress and maintaining accountability for commitments made by industry and government. The taskforce is expected to mobilise industry leadership and examine barriers at critical career transitions, including access to senior technical and leadership roles. These are complementary responsibilities, because sustained progress will require visible leadership, organisational commitment and the willingness to challenge established practices, supported by evidence about what those efforts actually achieve.
Meaningful accountability depends on understanding which interventions produce durable improvements, for whom and under what circumstances. That requires attention to participation and progression across educational and professional pathways, including retention at critical career transitions, access to continuing professional development, entry into senior technical roles, pay and promotion, and the experiences of women facing overlapping barriers associated with ethnicity, disability, socioeconomic background or geography. Aggregate workforce figures remain important, but they cannot explain those processes by themselves. Good evaluation needs longitudinal evidence, qualitative understanding of people’s experiences and transparent reporting of limitations, alongside a willingness to examine whether the underlying culture of an organisation is changing. Employers, educational institutions, government and professional bodies have different responsibilities within this system, but each should be able to explain what it is doing, why it expects those actions to work and what evidence would lead it to reconsider them.
Who gets to shape the digital society?
The significance of this extends well beyond employment and representation within a particular profession. The people designing and developing digital systems make judgements about the problems worth solving, the requirements to prioritise, the risks considered acceptable and the interests of those affected by their decisions. Some choices are explicit, while others become embedded in technical architectures, development practices and organisational assumptions, often long before the wider public encounters the resulting systems. The composition of the profession, and the distribution of authority within it, therefore have consequences for whose experiences inform technological development and which assumptions are open to challenge.
At stake is a question that connects computing education, professional practice and technology governance: what makes a good digital society, and who decides? There is no single or uncontested definition of the digital good, and decisions about it cannot legitimately be left to technology companies, professional experts or policymakers acting on their own. A society increasingly shaped by digital technologies needs meaningful opportunities for people to participate, exercise agency and influence decisions affecting their lives, supported by institutions capable of scrutiny, challenge and accountability. The contribution of engineers and computer scientists is essential within that wider system, but their professional judgement must remain connected to the interests and legitimate authority of the people and communities their technologies affect.
A broader range of experiences and perspectives within the computing profession can help identify neglected requirements, question established assumptions and expose risks that might otherwise receive insufficient attention. Diversity does not automatically guarantee reliable or equitable technology, any more than homogeneous teams necessarily produce poor systems, and public legitimacy cannot be secured through the composition of technical teams alone. These outcomes depend on engineering practice, institutional governance and the capacity for independent scrutiny, alongside the participation of people affected by technological decisions. Persistent exclusion nevertheless narrows the pool of expertise able to exercise technical judgement and can reinforce institutional blind spots, particularly where authority becomes concentrated within relatively homogeneous professional networks.
These concerns also arise in the debate about human rights and meaningful oversight in AI systems. As AI becomes more deeply embedded in consequential decisions, we need to pay attention to the people and institutions exercising technical authority, the perspectives informing their choices and the opportunities for those choices to be challenged. In a recent post, I argued that substantive human involvement depends on controllability, legitimate authority and contestability across the wider sociotechnical system. Questions about who develops those systems, whose expertise carries weight and how institutions distribute professional authority belong within the same discussion, because technical capability acquires much of its significance through the people and institutions empowered to use it.
Ada Lovelace Day earlier this month provided an opportunity to recognise the achievements of women in computing, just as the response to Margaret Hamilton’s death has reminded us of her extraordinary contribution to software engineering. The celebration and recognition of those achievements matter, particularly when histories of computing have too often overlooked or diminished the work of women, but there is an uncomfortable tension between recognising exceptional individuals and accepting professional cultures in which familiar barriers to participation and advancement remain. Their achievements should prompt us to examine the educational, professional and institutional conditions under which the next generation will develop and exercise its expertise.
Responsibility for changing those conditions cannot rest primarily with the women who experience the barriers, or with the relatively small number who have become visible advocates for change. Those of us who have benefited from the opportunities computing provides, particularly men in senior academic, engineering and professional positions, have a responsibility to examine how our own decisions affect the opportunities available to others. I include myself in that obligation, and the same scrutiny must apply to how I exercise my own professional authority.
The Royal Academy of Engineering’s work on majority allies and inclusive cultures recognises the importance of people in positions of influence helping to change organisational expectations and behaviour. In practice, that means examining whose contributions we acknowledge, whom we mentor and sponsor, how recruitment and promotion decisions are made, who is entrusted with technically demanding work and leadership responsibilities, and whether we are willing to challenge exclusionary behaviour even when doing so is inconvenient. Allyship has little substance if it remains detached from those everyday exercises of professional judgement and authority, and from a willingness to be accountable for their consequences.
Margaret Hamilton spent much of her career insisting that software engineering deserved the rigour, recognition and respect accorded to other engineering disciplines, while demonstrating what those principles meant through the design and development of systems on which human lives depended. The profession she helped establish now has an enormous influence over the technologies and institutions shaping contemporary society, and that influence brings responsibilities for how we educate, recognise, support and promote the people developing them. We should expect evidence of progress from our professional institutions, challenge practices that continue to exclude capable people, and be prepared to examine the consequences of our own decisions.
Hamilton helped establish the authority of software engineering by demonstrating the seriousness of its methods and responsibilities. The future of that profession deserves the same seriousness from all of us who have a part in shaping it.
