The George Washington University Open Source Program Office (OSPO) hosted its first annual conference on open source strategy, software and initiatives last week, with GW Provost Christopher Alan Bracey calling the gathering “an exciting milestone for GW.”
The GW OSPO conference welcomed students, faculty, researchers, industry leaders and open-source advocates to the two-day event, which came nearly two years after GW OSPO was created with a grant from the Alfred P. Sloan Foundation.
“Open source is more than just software, it’s a philosophy of collaboration, transparency and shared innovation,” Bracey told the conference. “Some of the most transformative technologies today, from artificial intelligence to cybersecurity, depend on open-source ecosystems.
“At GW, we embrace the principles of open education, open science and open data as a way to drive creative solutions to society’s most complex challenges,” Bracey said.
Day one of the conference was kicked off by keynote speaker Steve Crawford, a program executive from NASA’s Open Source Science Initiative. He discussed the history and importance of open science at NASA, and he highlighted the Astropy open source project, which had a start on a Python mailing list and now has been cited in over 10,000 publications. Inessa Pawson, open source program manager at OpenTeams, followed with her presentation.
GW OSPO Director David Lippert noted that the office is based under GW Libraries and Academic Innovation and supports all in the GW community interested in learning open science and open source best practices for sharing knowledge.
The OSPO services and initiatives include workshops, tutorials, student award programs, as well as collaborations with GW Coders and George Hacks to bring speakers to campus, he said. It is also partnering with government agencies and various companies to provide capstones and internships for GW students.
In opening remarks on the second day of the conference, Lorena Barba, GW OSPO faculty director, noted that the conference was being held at “a pivotal moment for higher education” as “universities across the nation face unprecedented challenges: funding constraints, questions about relevance and skepticism about the value we provide.”
“The Open-Source Program Office invites us to reimagine how academic knowledge is created, shared and sustained,” she said. “Yet the mission of the academy to create and share knowledge has never been more vital.”
Centers for Medicare and Medicaid Services
Barba introduced speakers from the U.S. Centers for Medicare and Medicaid Services (CMS): Andrea Fletcher, chief digital strategy officer, and Remy DeCausemaker, the open source lead for the digital service at CMS. They described projects at the agency that demonstrate open source strategies can cut costs, reduce burdens and risk.
CMS is 22% of the federal budget and supports health care to more than half of Americans.
Fletcher noted that when the CMS open source office began, leadership at the Department of Health and Human Services were concerned sharing technology opened the agency to risks from hackers. By contrast, they decided to pay hackers to identify weaknesses in CMS systems running the first bug bounty on federal health systems, in partnership with CISA.
“We found over 200 vulnerabilities and 18 critical vulnerabilities,” Fletcher said. In just four months, the effort alleviated $180 million in vulnerabilities. “We are demonstrating that our open source software is as secure if not more secure than some of the [software] that is not open source.”
Provost Christopher Alan Bracey called the GW Open Source Program Office first annual conference "an exciting milestone for GW." (Lily Speredelozzi/GW Today)
GW Computational Biology Institute
The conference capstone featured Keith Crandall, professor and founding director of the GW Computational Biology Institute (CBI) who has put GW on the map in the world of bioinformatics.
Crandall can claim more than 20,000 citations for the ModelTest software that is used to sequence DNA, which he downplayed but said was useful 20 years ago when most biologists didn’t know much about programming or creating software.
Crandall’s presentation, titled “From Open Source to Open Science,” addressed CBI’s development of software and implementation and sharing of data.
Open source software is deeply connected to CBI’s investigation into genetics, natural history, and human health and science.
“All of our software is out there. Anybody can use them,” Crandall said, including private industry. “But that’s what happens with academic software. Academic software is more vetted and more recent than whatever is plugged into commercial software, which is older, and they never change it.”
He shared the story of how CBI’s ambitious Earth BioGenome Project, which aims to sequence the genome for every living thing, grew out of a group experience he had at a National Science Foundation Ideas Lab decades ago when they came up with the idea to build a Tree of Life, now the Open Tree of Life to which the BioGenome Project contributes genetic data.
“Any moment you can download an updated Tree of Life. And it is super useful,” he said. “You have to go all over the place because there are funny little groups of taxonomists in funny pockets of the world doing wonderful things with their taxonomy.”
The data is linked to GenBank at the National Library of Science where, he said, CBI has been “sharing from day one. Sharing genetic data for us is trivial. It is absolutely the standard. You cannot publish a paper without sharing your genetic data.”
But there is a hitch. Metadata, research based on that data, particularly human health research, is not so openly available. Access to that information must be approved by the original research institution, the National Institutes of Health (NIH) and the National Cancer Institute (NCI).
“You cannot download it,” Crandall said. “You have to go there to do the work.”
In some instances, you must pay for access to data from some research projects, he said. Compounding that are costs associated with some of the peer review journals that publish the research.
“Some journals have open access; some have processing fees. We need to figure this out because it is not healthy for the university at all.”
“For an academic, the ecosystem is not just about the software,” Crandall said. “Certain data is more well positioned for open access, like genetic data. Others not so much, even metadata for most studies. And in publishing we have a long way to go.”
The conference roster of speakers featured GW teams of professors and students who showcased their development of open-source applications, and representatives from both industry and other federal agencies.