Open Source Voting Systems Resolution #2

Adopted by the San Francisco Elections Commission (6-0) on June 20, 2018.

[The Commission is adopting this resolution to supplement (and not replace) its first Open Source Voting Systems Resolution, adopted on November 18, 2015, since many events have transpired since that time.]

	WHEREAS, The San Francisco Elections Commission (“Elections Commission”) on November 18, 2015 unanimously passed an Open Source Voting Systems Resolution that, in part—
(a) Encouraged the Mayor and Board of Supervisors to initiate and fund a project to develop and certify an open source voting system; and
(b) Outlined several key project characteristics, and encouraged the Mayor and Board of Supervisors to consider incorporating them; and
(c) Established a policy for the San Francisco Department of Elections to support and work towards the adoption of an open source voting system;
	WHEREAS, Mayor Ed Lee in the 2016 budget process allocated $300,000 towards a planning and assessment phase for the project, approximately half of which was allocated for a “business case” to assess the feasibility of the project, including informing the City and County of San Francisco (“San Francisco”) of the associated costs, timelines, options, and possible next steps;
	WHEREAS, The Elections Commission on April 19, 2017 established an Open Source Voting System Technical Advisory Committee (“TAC”) to provide technical guidance, ideas, and support to the Elections Commission on ways to improve and help ensure the success of San Francisco's open source voting system project;
	WHEREAS, The Department of Elections in May 2017 issued the RFP for the business case, which required that, “Any Contractor who prepares or assists in preparing the business case described in this RFP will not be eligible to perform subsequent services for the City in relation to the possible development of an accessible, open source voting system,” which resulted in organizations with the most interest and expertise not bidding;
	WHEREAS, The winning bidder, Slalom, whose contract was signed September 29, 2017, had no prior experience in elections or voting systems and little or no experience with open source development efforts;
	WHEREAS, Travis County, Texas on September 28, 2017 cancelled its STAR‑Vote project, in large part because its monolithic, waterfall-like approach sought to build all components at once as opposed to developing components delivering usable functionality incrementally, and which differed from San Francisco’s project because it—
(a) Was larger in scope to the extent that its core purpose was to fundamentally change how voting was performed, recorded, tabulated, and verified;
(b) Was based on a proprietary source rather than an open source development plan and only anticipated becoming open source eventually; and
(c) Lacked a business / development plan to incrementally provide usable functionality within the available funding capacity;
	WHEREAS, Slalom delivered its final report to San Francisco for review by the Mayor’s Office and San Francisco’s Committee on Information Technology (“COIT”) on March 14, 2018, more than six weeks after the January 26, 2018 due date and well into San Francisco's annual budget process;
	WHEREAS, Slalom's report recommended a waterfall-like approach and lacked the information, specificity, and credibility needed for San Francisco leaders and stakeholders to decide on next steps, including—
(a) Recommended a second, much more expensive planning phase of $1.1 to $1.2 million that the report called a “Discovery Phase,” which would take place before doing any actual development and whose purpose included doing many of the things the business case was supposed to do;
(b) Failing to consult with vendors and organizations that the Commission invited to present at its October 21, 2015 meeting—subsequent to the Department of Elections’ August 6, 2015 RFI (REG 2015-01) in relation to obtaining a new voting system—all of whom presented cost estimates much lower than those provided in Slalom’s report;
(c) Provided excessively high cost estimates with extremely wide ranges, including—
(1) $850,000 to $2 million to develop a “Vote Tabulator System,” a simple component responsible for summing vote totals;
(2) $860,000 to $2 million to develop a “Vote Reporting System,” a simple component responsible for generating static reports of vote totals; and
(3) $500,000 to $1 million per year for “cloud-based hosting,” even though only a remote accessible vote by mail component is allowed to be connected to the internet, such a component would require a much smaller hosting requirement, and such a component is not a requirement to implement any other part of the system;
	WHEREAS, TAC on March 8, 2018 voted unanimously to recommend to the Elections Commission that—
(a) “the Mayor budget for the Department of Elections a full-time staff person who can serve as the project lead / product owner for the project and report directly to the Director of Elections”; and that
(b) “an agile, incremental approach be taken towards the project, in which components can be piloted and used in real elections [in conjunction with the proprietary system in use] as the components are developed and certified,” starting with the components needed to tabulate vote-by-mail ballots;
	WHEREAS, The Elections Commission on March 21, 2018 voted unanimously to approve TAC's March 8, 2018 recommendations;
	WHEREAS, TAC on June 14, 2018 voted unanimously to adopt a statement on Slalom’s “Open Source Voting System Feasibility Assessment,” beginning, “Certain inaccuracies and basic omissions in Slalom’s Report lead us to seriously question the report as a whole. We recommend not relying on it for determining the future direction of the project”;
	WHEREAS, TAC also on June 14, 2018 voted unanimously to adopt a project to implement a full-featured election results reporter as a proof of concept, to show in part that a reporting component can be developed for much less cost than the $860,000 to $2 million estimated by Slalom;
	WHEREAS, The Open Source Voting System Resolution unanimously passed by the Elections Commission on November 18, 2015 encouraged that the project “First, hire a project director with technical expertise to be responsible for planning and leading the project, including working with stakeholders, collaborators, and regulators; drafting system requirements; and selecting and managing technical contractors, as necessary”;
	WHEREAS, The Department of Technology and Department of Elections jointly submitted to COIT an application for funding the open source voting system project in FY2018-19 with an additional $960,000; and in response COIT recommended only an additional $300,000, for a total of $425,000 available for FY2018-19;
	WHEREAS, San Francisco’s Department of Human Resources (“DHR”) in 2017 started a technology project to modernize its hiring practices (the “Hiring Modernization Project”) using an iterative, modular approach that internally it estimated could cost between $20-30 million over five years;
	WHEREAS, COIT on May 4, 2018 recommended allocating $1 million in seed money for the first year (FY2018-19) of the Hiring Modernization Project, for the purposes of hiring a Project Manager and starting incremental development—even though DHR recognizes that their overall estimate will continue to evolve since the project is being approached in a modular and iterative way;
	WHEREAS, The Elections Commission is responsible for the proper administration of the general practices of the Department of Elections, which in turn is vested with the conduct and management of matters pertaining to elections in San Francisco;
	WHEREAS, Since voting systems pertain to and are used in elections in San Francisco, their development by San Francisco should require the same level of transparency and public oversight that elections themselves require;
	WHEREAS, Progress towards planning and development of an open source voting system has so far been hampered by the lack of both clear designations of accountability and a person with appropriate skills and experience to lead the project;
	WHEREAS, Significant efforts and progress have been made by the California Clean Money Campaign in conjunction with some San Francisco elected officials to provide for state matching funds for open source voting systems;
	WHEREAS, Clearer and renewed commitments from the Mayor and the Board of Supervisors for an open source voting system, including a commitment for greater funding in FY2018-19, would strengthen the project’s ability to attract effective development collaborators and supplementary sources of funding;
	WHEREAS, Board of Supervisors President and Mayor-elect London Breed, Board of Supervisors Budget and Finance Committee Chair Malia Cohen, and Board of Supervisors Member Jane Kim recently made public statements in support of effective funding to start developing an open source voting system project, in response to the Elections Commission’s request of $4 million for FY2018-19; now, therefore be it
	RESOLVED, That the Elections Commission calls on the Mayor and Board of Supervisors to state their commitment to effectively proceeding with San Francisco’s open source voting system project, so that the Elections Commission, Department of Elections, and the rest of San Francisco can state publicly and unambiguously that San Francisco has decided to move forward, which will allow San Francisco to more effectively identify and attract potential collaborators and other funders, including the State of California, nonprofits, and other counties; and, be it
	FURTHER RESOLVED, That the Commission ask the Mayor and Board of Supervisors that the project be funded and developed incrementally, so that San Francisco can support a robust development schedule in FY2018-19 that—
(a) Develops and uses some open source voting system components sooner and more cost-effectively, without necessarily having to first fund the entire project;
(b) Develop cost estimates and project management approaches for later stages of the project by learning from real-world experience, as opposed to planning abstractly on paper; and
(c) More effectively attract and recruit project collaborators, by being able to show real progress using innovative approaches; and, be it
	FURTHER RESOLVED, That the Elections Commission ask the Mayor and Board of Supervisors to name the Department of Elections as the owner of the project to develop and certify an open source voting system, so that the Elections Commission and the public can, through the authority and meetings of the Elections Commission, have oversight over and regular visibility into this fundamental matter pertaining to elections; and, be it
	FURTHER RESOLVED, That it is the policy of the Elections Commission that the open source voting system project should be led by a dedicated project director / project manager who reports to the Director of Elections; who has experience in agile methods, open source development, elections and voting systems, and the management of technical projects; and who will—
(a) Work with the Director and other stakeholders in San Francisco;
(b) Use TAC and other Departments among their resources, including learning from the model of DHR's Hiring Modernization Project;
(c) Use the RFI process to learn possible approaches from vendors; and, be it
	FURTHER RESOLVED, That the Elections Commission rejects the cost estimates in the Slalom Report as poorly supported and not credible, and the waterfall approach recommended in the report as outmoded, inconsistent with the agile approach preferred by the Commission, and generally not in San Francisco's interest due to the increased cost and risk; and, be it
	FINALLY RESOLVED, That the Commission encourages the Department of Elections to continue its best practices of working collaboratively with other Departments, including the Department of Technology, San Francisco Digital Services, the Office of Contract Administration, the Mayor's Office, and other members of COIT on the project.