Forum Thread Archive Note, November 2004

Archive record — this URL held a discussion thread posted in November 2004. The thread is not reproduced; see the note below.

Archival note

This URL held a thread in the National Ballot Integrity Project's breaking-news discussion board, posted in November 2004. The thread consisted of a post by a forum member reproducing an article from an outside publication concerning the corporate history of a company that supplied voting equipment.

The thread is not reproduced in this archive, and neither its content nor its claims are repeated, paraphrased or endorsed here. There are three separate reasons. The post consisted almost entirely of copyrighted third-party material republished without permission. It contained unverified allegations about a named company and named individuals, which this archive does not repeat in any form. And it was posted under a private individual's chosen forum name, which is not carried forward. The general policy on the forum material is set out on the forum archive page.

The URL is preserved because it has been cited elsewhere for two decades and a reader following such a citation deserves an honest account of what was here and why it is not. What follows is a factual explanation of the underlying subject the thread touched on — how voting equipment is tested and certified in the United States — drawn from current official sources.

What certification means

A voting system is not a single device. It is a combination of hardware, firmware, software and documentation: the equipment a voter interacts with, the scanners or tabulators that read votes, the software that defines a ballot and accumulates results, and the procedures for operating all of it. Certification is a formal examination of that combination as a whole against a published standard, resulting in a decision that the configuration tested does or does not conform.

Certification is not a warranty that a system is secure in the abstract, nor that it will be used correctly. It establishes that a specific, versioned configuration was tested against defined requirements by a qualified party, and that the test results are on the record.

The federal process

At federal level the U.S. Election Assistance Commission operates a testing and certification programme, and its structure is published in full as the system certification process. In outline it works as follows.

• A published standard. Systems are tested against the Voluntary Voting System Guidelines, which set requirements for functional capability, security, auditability, usability and accessibility. Successive versions of the guidelines have raised those requirements considerably since the first version was adopted in 2005.
• Accredited testing laboratories. Testing is performed by voting system test laboratories that are themselves accredited, with the accreditation process resting on technical criteria developed with the National Institute of Standards and Technology through its voting programme. The manufacturer does not test its own system.
• A registered configuration. Certification attaches to a precise version. Changes to hardware or software require either a modification review or a new certification, so a system in the field can be matched against a documented tested configuration.
• A public record. Certified systems, their test reports and their modification histories are published in the commission's register of certified voting systems. This is the single most useful reference for anyone wanting to know what has actually been examined and what the examination found.

The state layer

Federal certification does not, on its own, authorise the use of any system anywhere. Election administration is a state responsibility, and each state decides what to require. Some require full federal certification; some require testing by an accredited laboratory without requiring the federal certificate; some run their own state certification process, sometimes considerably more demanding than the federal one, including state-specific functional testing and field trials. A handful require little formal certification at all.

Layered on top of certification, and distinct from it, is pre-election logic and accuracy testing: the local process by which a county tests its own configured equipment, with its own ballot definitions, before each election. Certification asks whether a system design conforms to a standard; logic and accuracy testing asks whether this machine, loaded with this ballot for this election, counts correctly. Both are necessary, and the 2004 material on this site was principally concerned with the second — see the accountability procedures page.

How this differs from the 2004 picture

The relevance to this archive is that essentially none of the above existed in operational form when the original thread was posted. In November 2004 the Election Assistance Commission was under a year old, the guidelines had not yet been adopted, the accredited laboratory programme had not been established, and there was no published federal register of certified systems and test reports. Judgements about equipment were made against older standards administered under a different arrangement, with results that were not systematically published.

That absence is the context in which citizen groups conducted their own research into equipment and vendors, sometimes carefully and sometimes not, and it is the context for the 2004 argument recorded on the statement of purpose. It is also why this page substitutes verifiable current sources for a two-decade-old forum post: the questions the post was reaching for now have documented, publicly checkable answers. The sequence in which those mechanisms were built is set out on the timeline page.