Over the past few years, the Software Heritage Symposium has evolved into a cornerstone event that unites diverse stakeholders from different sectors. The momentum began with the inaugural Symposium in 2021, which celebrated Software Heritage’s 5th anniversary. The momentum continued with a second Symposium in 2023, further solidifying the importance of our mission. Most recently, in February 2024, we convened the 3rd Symposium, reinforcing our commitment to safeguarding the vast knowledge embedded in software source code.

Organized in collaboration with UNESCO, the event provided an opportunity for attendees to explore the varied impacts and applications of software source code archival across different sectors. By bringing together a diverse group of presenters and panelists, the event fostered a rich exchange of ideas and experiences, highlighting the ongoing significance of software source code in society and the importance of preserving our digital legacy. Here, we delve into the key insights and discussions that emerged during this enlightening event.

UNESCO and Inria’s commitment to Software Preservation

Launched in 2016 by Inria, Software Heritage, is a pioneering initiative to collect, preserve, and share all software source code as a common infrastructure at the service of cultural heritage, science, industry, and society as a whole.

In a significant milestone, UNESCO and Inria joined forces in 2017, with the formal partnership agreement, to collaborate on preserving and disseminating software source code knowledge. This partnership reflects a commitment to safeguarding our digital legacy for future generations, enhancing access to information, and supporting global innovation and education.

Symposium Highlights

UNESCO – Paris | © Inria / Photo M. Magnin

The welcome address and opening of the third annual Symposium was launched by Mr Fackson Banda, Chief of the Unit for Documentary Heritage at UNESCO, whose Documentary Heritage Unit is the Secretariat of the Member of the World (MoW) programme, co-host of the event.

In her opening remarks, Ms Marielza Oliveira, Director Communications and Information – Division for Digital Inclusion, Policies and Transformation, UNESCO, emphasized the pivotal role in preserving and sharing the invaluable cultural heritage embedded in software code, «software is the lifeblood of our digital age permeating every aspect of our lives and driving innovation across industries»

Marielza Oliveira at UNESCO – Paris | © Inria / Photo M. Magnin.

Mr Gilles Mathieu, representative, French Ministry of Research and Higher Education, highlighted the ministry’s commitment to open science by advocating for the promotion and preservation of research-generated source code through archiving in Software Heritage.

Roberto Di Cosmo at UNESCO – Paris | © Inria / Photo M. Magnin

Mr Jean-Frédéric Gerbeau, Inria Deputy CEO for Science, reminded the challenges that Software Heritage met in harvesting the software source code, as he puts, arguably more difficult operation than referencing the web itself. Recognized that the Software Heritage archive has truly flourished, emphasizing the recent projects that pave the way to new opportunities.

Reinforcing the role of software and its source code being a fundamental enabler in all human activities and setting the scope for the symposium, Mr Roberto Di Cosmo, Director of Software Heritage, stated:

“Software Heritage is building a universal source code archive as one infrastructure, shared and mutualized across all fields of endeavour, because software is the digital fabric that binds all them together”

Following the welcome address, the main topics were discussed in panels and presentations:

• Empowering Innovation through Software Source Code in Industry and Governments
• Scientific Challenges in Analyzing and Learning from Preserved Source Code
• Software Source Code in the Open Science Ecosystem
• Software Source Code as Documentary Heritage

Industry & Public Administration Panel

Kate Stewart at UNESCO – Paris | © Inria / Photo M. Magnin

The first panel introduced topics from industry and public administration and Industry, moderated by Mr Roberto Di Cosmo and including: Mr Guillaume Avrin, National coordinator for artificial intelligence, Direction Générale des Entreprises, Mr Omar Mohsine, Office of the United Nations Special Envoy on Technology, Mr Marc Palazon, Board member and President of the Open Source commission, Numeum and Ms Kate Stewart, VP Dependable Embedded Systems, Linux Foundation.

Ms Kate Stewart emphasized the importance of reproducibility and the need of an infrastructure for tracking the source of truth, where Software Heritage and the SWHID play a key role. Mr Omar Mohsine shared that the UN engagement toward Open Source by building upon a strategy composed by three pillars: policy, culture change and the Open Source heroes. In France, the public administration understood that Open Source is key and today the biggest market in Europe for Open Source is in France, stated Mr Marc Palazon.

Mr Guillaume Avrin discussed the national strategy for AI and the announcement in 2023 of the extension of the super-computing facility on the Plateau de Saclay.

Guillaume Avrin, Roberto Di Cosmo, Omar Mohsine, Marc Palazon et Kate Stewart at UNESCO – Paris |© Inria / Photo M. Magnin

Scientific Challenges: Rust Analytics for Software Heritage

Sebastiano Vigna at UNESCO – Paris | © Inria / Photo M. Magnin

Mr Sebastiano Vigna presented Rust Analytics for Software Heritage, as well as the graph representation approach with the WebGraph framework. Showing the  results of a collaboration between Télécom Paris and the Università degli Studi di Milano, on a new graph compression framework written in Rust on the Software Heritage history graph. This innovative approach for representing large graphs opened new possibilities with a more predictable performance, and three times faster!  Preparing for the future growth of Software Heritage.

Scientific Challenges: Big Code

Landro v. Werra, Hugging Face and Harm de Vries, Staff research scientist, ServiceNow Paris | © Inria / Photo M. Magnin

Mr Leandro von Werra and Mr Harm de Vries offered insights into the development of large language models for code during their presentation. They showed how the BigCode project that they lead builds the most open and transparent models available today, making available all the data collection, filtering and training pipeline, and providing tools for developers to check whether their code is in the training dataset, to support opt out. They detailed the reasons for establishing a collaboration with Software Heritage: a shared engagement to contribute to a common good. «We partner with Software Heritage  to ensure that the source code used to build the models is accessible and identifiable, enhancing the transparency of our efforts.»

Open Science Panel

The following session included lively discussions by the Open Science Panel, which demonstrated how software source code is not only a tool for the preservation of the world’s software heritage but also an instrument at the service of Open Science emphasizing the importance of Open Source as an enabler for Open Science.

Mr Christopher S. Marcum, Senior Statistician and Senior Science Policy Analyst at the Office of the Chief Statistician of the United

States, shed light on the substantial source code repository maintained by the U.S. Bureau of Statistics (USB) on GitHub. «The big thing here that is relevant for Software Heritage is that by U.S. federal policy, federal agents are required to share their open-source code…» noting that it is a policy of the United States government for federal agencies to share, at a minimum, the metadata associated with their code repositories on code.gov, aligning with executive orders previously mentioned.

Ms Claudia Bauzer Medeiros, Professor at the University of Campinas (UNICAMP), declared that «Software Heritage is a treasure,» while quoting Paul Valéry

It depends on those who pass

Whether I am a tomb or treasure

Claudia Bauzer Medeiros at UNESCO – Paris |© Inria / Photo M. Magnin

Whether I speak  or am silent

The choice is yours alone.

~ Paul Valéry

Reminding us that «unless you understand its nature and reuse it, unless you take advantage of this treasure, we must make it abundantly clear that it’s not a mere repository to be buried as a tomb.» Claudia’ words highlighted the importance of recognizing and actively engaging with the wealth of knowledge preserved within Software Heritage, emphasizing its potential to drive innovation and progress in the digital age.

Katlin Thaney at UNESCO – Paris | © Inria / Photo M. Magnin

Ms Kaitlin Thaney, Executive Director of Invest in Open Infrastructure, encouraged attendees to broaden their perspective on open research during her address, emphasizing the synergies between open research and other areas of innovation. «If you rely on any form of open research in your work in research and development, it’s part of your process,» she asserted. Kaitlin urged participants to consider not only the systems they utilize but also to explore open alternatives and ways to give back. She emphasized the importance of collaboration beyond government and philanthropic support, as they alone are unable to sustain long-term efforts. Kaitlin extended her gratitude and congratulations to Software Heritage for their leadership in this endeavor, concluding her remarks with an invitation for attendees to reflect on these principles.

Mr Bhanu Neupane, Programme Manager for ICT and Sciences and Open Access to Scientific Research at UNESCO, highlighted the need for reliable indicators to gauge the impact of free and open-source software on driving the open science agenda globally. «It’s crucial to develop indicators that member states can use to measure the extent to which free and open-source software contributes to scientific research in their countries».

The Open Science panel at UNESCO – Paris | © Inria / Photo M. Magnin

Mr Roberto Di Cosmo, CEO of Software Heritage, underscored the universal nature of software during his remarks. «Software is designed for all,» he said. «Not just for specific countries, research areas, or industries, but as a common infrastructure for humanity.» Roberto stressed the importance of communication and collaboration with diverse communities, acknowledging the complexity of the open-source ecosystem. He challenged the notion that open source alone guarantees value, highlighting the need for quality control and decision-making in software development and funding. Roberto outlined Software Heritage’s mission to provide a centralized infrastructure accessible to all, aiming to streamline the multitude of individual repositories into a unified resource for the future.

Software Source code as part of Memory of the World Panel

The last session was a Panel on Software Source code as part of Memory of the World moderated by Mr Fackson Banda, which addressed different aspects of digital cultural heritage and the place of source code the larger ecosystem of cultural heritage preservation and accessibility.

Rosana Lanzelotte at UNESCO – Paris | © Inria / Photo M. Magnin

Ms Rosana Lanzelotte, President of Musica Brasilis, showcased their collection of 6283 free sheet music downloads of Brazilian music scores, emphasizing the need for Music Character Recognition (MCR) software to decode handwritten scores. Aligning with FBG 11.4, Musica Brasilis contributes to preserving cultural heritage by going digital and promoting accessibility while adhering to the FAIR principles. Their collaboration with IICT for digital preservation and interoperable metadata exchange underscores their commitment to open access. Rosana highlighted the adaptability of Musica Brasilis’s web software, which can now support initiatives beyond its initial scope, promoting cooperation and potential contributions to networks like Software Heritage.

Mr Pio Pellizzari, Delegate of IASA, emphasized the importance of digital archiving for preserving access to invaluable audiovisual materials, highlighting the need for comprehensive documentation and collaboration to address evolving preservation needs effectively.

Valérie Schaferat UNESCO – Paris | © Inria / Photo M. Magnin

Ms Valérie Schafer, Professor at C2DH – University of Luxembourg, expressed her enthusiasm for the Software Heritage mission, highlighting its transparent technical processes and the open large community. Furthermore, Valérie quoted Lawrence Lessig’s statement «code is law» and shared the broader implications of code as a cultural artifact, shaping societal discussions around politics, gender, and ideology. She emphasized the need to engage citizens in understanding coding basics and its societal impact, envisioning a future where coding literacy becomes more widespread. Reflecting on her own research in web archives, Valérie explored the potential of hidden layers in archived web pages and praised Software Heritage’s efforts to narrate the history of code, acknowledging the diverse range of contributions from scholars in the field.

Conclusion

UNESCO – Paris | © Inria / Photo M. Magnin.

The Software Heritage 2024 Symposium served as a testament to the collective commitment to preserving our software commons. As we embark on the journey ahead, we would like to take this opportunity to thank UNESCO for their continuous partnership, our sponsors for their support and our community for their engagement. Let us continue to nurture our digital heritage while building a large community, ensuring that future generations have access to the universal source code archive.

Didn’t make it to the event? No worries!

Discover the slides to catch up on the presentations, view the snapshots from the day on our event webpage, and watch the sessions online.

Visit our Symposium 2024 webpage!

Software Heritage is a universal software source code archive project, whose aim is to recover, preserve for the very long term and share all publicly available source code, together with its development history (e.g., as stored in version control systems). The Software Heritage archive already contains over 17 billion unique source files and 3.6 billion commits, retrieved from over 266 million software development projects. The Software Heritage initiative, hosted by the Inria Foundation, is an entirely free software (FOSS) and non-profit project.

The Position

We are looking for an experienced Big Data-oriented software engineer. The ideal candidate will have significant interest and experience in large-scale data processing and exploitation architectures, including storage, indexing and retrieval.

You can consult a more detailed list of our current projects on the Software Heritage Roadmap 2023 (https://docs.softwareheritage.org/devel/roadmap/roadmap-2023.html)

Main tasks and activities

– Setting up a data processing architecture (a la Spark)
– Design and modeling of Big Data architectures
– Implementation of solutions based on defined architectures
– Set up Big Data pipelines

Skills

The ideal candidate will have experience in Big Data development and architecture, preferably in an open-source context. We expect self-organization and autonomy skills commensurate with the candidate’s experience. Participation in existing FOSS projects in any capacity (developer, community organizer, technical writer, etc.) is an added advantage.

The following skills are expected:

– Mastery of a large-scale data processing system (e.g. Apache Spark, Flink, or Hadoop)
– Fluent software development skills (basics in Rust and Python)
– Good level of English (written and spoken)
– Use of Git
– Use of continuous integration tools (e.g. Gitlab and/or Jenkins)

Knowledge and experience of the following will be considered an asset:

– Experience in data processing on a scale of tens of terabytes or even petabytes
– Experience with Cassandra and Kafka
– Knowledge of Java
– Knowledge of Kubernetes
– Data visualization

Software Heritage is a complex technical architecture, based on many different technologies, which continues to evolve. We do not expect candidates to master all of them, but rather to be open to discovery and learning. Prior knowledge of one or more of the above-mentioned subjects will help in the process of getting to grips with the project, but we encourage you to apply whatever your level of experience in these technologies.

Working conditions

We are a team of 15 people, including 9 technical staff (5 developers and 4 sysadmins).
Autonomy, transparency and consultation are at the heart of our values (the project is free and open source).

Most of the team is based at the Inria center in Paris, but the position is open to any location in France close to an Inria center (Bordeaux, Lille, Lyon, Grenoble, Rennes, Saclay, Sofia Antipolis, Nancy).

The contract offered by Inria is a 2-year renewable full-time fixed-term contract, with the prospect of a permanent position.
– Telecommuting: 90 days/year (average 2 days per week)
– Vacation: 35 days + 10 days RTT
– Salary range: 30 to 70 k€ depending on profile and experience.

Application

Please send your application (CV + cover letter) to hiring@softwareheritage.org

Julian Coccia, CTO, SCANOSS

— Nicolas Fressengeas,
Vicepresidente de la Université de Loraine
encargado de Datos Digitales y Ciencia Abierta

Emmanuel Chiva,
Delegado General de Armamento

Challenges and opportunities

The advent of Large Language Models (LLMs) capable of generating code presents both a challenge and an opportunity. The challenge lies in navigating the complex legal and ethical landscapes governing these innovations. To address this, we unveiled our statement on Large Language Models for Code in October 2023, outlining our guiding principles: openness, transparency, and respect for the authors.

The opportunity arises from the potential to make the vast body of knowledge embedded in humankind’s source code more accessible and reusable for a much broader community: this aligns perfectly with our core mission.

A first milestone with StarCoder2

Today, we are thrilled to see a first realization of this opportunity with the introduction of StarCoder2, the first-ever AI model for code developed using the comprehensive source code repository of the Software Heritage archive, and fully aligned with our principles for LLMs for code.

We congratulate Hugging Face, ServiceNow, and NVIDIA for their collaborative efforts to reach this important milestone within the BigCode project, showing commitment to ethical AI development, and advancing technology for the greater good. StarCoder2 is a significant step towards a world where the barriers to software development are lowered, where innovation is fueled by ethically developed AI, and where the digital commons serve as a foundation for future breakthroughs, ensuring that the knowledge derived from decades of software development benefits humanity at large.

Forward Together

As we celebrate this milestone, we are reminded of the ongoing importance of our mission to collect, preserve, and share the wealth of human knowledge embedded in software.

We invite the global community to join us in this exciting new chapter.

Software Heritage is more than just a repository; it signifies a commitment to a future where every line of code adds to our shared legacy and collective advancement. Embark on this journey with us—because every line of code, and every step towards ethical AI, counts.

— Thomas Wolf,
cofundador
Hugging Face

— Nicolas Chapados
Vicepresidente de Investigación,
ServiceNow

— Antoine Petit, Chairman and CEO, CNRS

The FAIR-IMPACT project launched in June 2022, has the role to support and disseminate FAIR-enabling practices, tools and services across scientific communities at a European, national and international level. Software Heritage’s role in the project is to establish guidelines for the collection and the curation of metadata to archive, reference, describe and cite research software. In June 2023, the Research Software MetaData (RSMD) Guidelines were published on Zenodo following a community workshop in March and a community review webinar in May.

Understanding the RSMD Guidelines

The RSMD Guidelines, written by Task 4.3 in the FAIR-IMPACT project, provide a comprehensive framework, offering flexible and adaptable recommendations for end-users across various disciplines and software development contexts. This deliverable encompasses:

• Introduction of goals, methodology, and use cases
• State-of-the-art review of existing practices and guidelines
• Comprehensive analysis of the metadata landscape
• Proposal for RSMD guidelines to collect and curate research software metadata
• A clear checklist for researchers

A living version of the document version will be available on the RSMD Guidelines repository.

Moving toward adoption

The journey doesn’t end with the proposal. Task 4.3 is committed to making the RSMD Guidelines normative within the academic community. This involves ongoing engagement with stakeholders, gathering feedback, and incorporating best practices and advancements in metadata management. By establishing these guidelines as a norm, the aim is to promote widespread adoption and adherence, leading to greater standardization and harmonization of metadata practices across research domains.

We offer two support actions that are designed to enhance the FAIRness and impact of research software:

• Path I: Assessing and improving existing research software using a new extension of F-UJI which implements some of the metrics for automated FAIR software assessment. Successful applicants to this support action will receive 4000 € to support their participation between May-September 2024.
• Path II: Implementing the Research Software MetaData guidelines for better archiving, referencing, describing, and citing research software artefacts. Successful applicants to this support action will receive 6000 € to support their participation between May-September 2024.

How can I implement the RSMD guidelines in the FAIR-IMPACT Open Call?

This goal can be achieved in many forms with the high-level purpose to better archive, reference, describe and cite research software artifacts.

Participants in path II of the support action will suggest implementation activities in their own resources or contributions to Open Source / Open Science existing projects, such as CodeMeta.

Mentored by software metadata experts, participants will strive to implement the RSMD guidelines, throughout a month-long challenge and will complete it by writing  a detailed implementation story to showcase the adoption of the RSMD guidelines. The outputs will be released in open access via Zenodo.

This support action will consist of four virtual workshops.

• Introductory session to provide context and background and to introduce scholarly infrastructures and tools to make Research Software a first class output (May);
• Introduction to the RSMD guidelines and examples of adoption (May);
• One day sprint to progress planned implementation activity (June);
• Post-assessment workshop where the Implementation story will be presented (September).

How to submit your proposal?

Apply on the Open Call page, after creating an account, you’ll find the form on the grants application dashboard. applications should include:

• Your name, contact details and organisational affiliation. Where the application is being made on behalf of a group, a nominated lead participant will need to provide these details;
• The type of organisation you are based at; The country you are based in (Please note that applicants must work in a European Union or Associated Country for the duration of the grant. For a full list of Associated Countries see here);
• Proposal for implementing the RSMD guidelines (Max. 500 words): What are the objectives of your proposed Research Software MetaData (RSMD) implementation? Additionally, specify how you plan to achieve this objective. A few examples are listed below
  • CodeMeta mapping tools;
  • Import/export tools in a software infrastructure;
  • UI improvements to facilitate software submissions;
  • SWHID exposure on software record;
  • Institutional documentation based on the RSMD guidelines;
  • Contribution to the existing CodeMeta generator or to the CodeMeta community.
    • contributing to the crosswalks
    • contributing to the documentation and to the user/developer guides
• Impact, adoption and dissemination
  • Describe how your participation in this support action will impact your work/project.
  • Describe how participating in the support action will support the wider adoption and uptake in your field of the tools, methods, and/or solutions employed.
  • Describe how you will disseminate the results of your participation in this support action.

Conclusion

The FAIR-IMPACT open call, with its focus on enhancing research software FAIRness and implementing metadata guidelines, presents a significant opportunity for collaboration and advancement. Together, let’s adopt and implement the RSMD Guidelines and work towards a future where research software is acknowledged as a first class output.

For further inquiries about the RSMD Guidelines or the FAIR-IMPACT project, feel free to reach out to: opencalls[at]fair-impact.eu

Your contributions are invaluable as we strive to build a more open and curated research software landscape.